System Guide

Welcome to Cowboy CRM

This guide is the complete operator's manual for Cowboy CRM — the autonomous AI brain running your call center. Every module, every button, every field, every workflow is documented here.

14 modules fully documented
11 real roles documented
Video walkthroughs on YouTube
PCT-patented AI engine

What is Cowboy CRM?

Cowboy CRM is an autonomous AI call center management system. It was built to eliminate manual decisions from call center operations — routing leads, building shifts, assigning agents, scheduling calls, and orchestrating communications — without a manager making individual decisions.

The system started development in 2017 under the name Astraq CRM, was later rebranded to EvolutionCRM.ai, and is now deployed as Cowboy CRM at cowyboymanager.com. It was built across 4 countries (Israel, Ukraine, Bulgaria, India) over 3 development phases, and currently serves 16 enterprise clients with 168,000+ client records in the database.

The Core PrincipleThe system makes decisions. You watch results. Cowboy CRM routes every lead, builds every shift, schedules every call — automatically. Managers intervene only for strategic decisions, not operational ones.

How to use this guide

Use the sidebar on the left to navigate between modules. Each module page includes: a description of what the module does, a screen walkthrough with annotated buttons, a field reference table, workflows for common tasks, and a role permission matrix showing who can do what.

Video walkthrough availableEvery major feature was recorded as a video walkthrough on the EvolutionCRM YouTube channel. Watch the videos alongside this guide for the fastest learning experience.

The AI Engine — How decisions are made

Cowboy CRM's core is a scoring algorithm that processes every lead across 40+ parameters in under 300ms. This algorithm is protected under PCT (Patent Cooperation Treaty) international patent and is also the engine behind SortExpress.com — the lead generation platform that feeds directly into Cowboy CRM.

1
Lead arrives

From any source: web form, affiliate, broker API, inbound call, manual entry, or SortExpress feed.

2
AI scores it

Geography, language, campaign, time of day, source quality, broker, device, historical behaviour — all scored simultaneously.

3
Agent matched

System selects the optimal agent based on language match, current load, shift, and composite performance score.

4
Action triggered

Call scheduled, email dispatched, or SMS sent — at the statistically optimal moment for that lead segment.

5
Outcome logged & model updated

Every result (converted, rejected, no-answer) feeds back into the scoring model. The brain improves with every interaction.

Getting Started

User Types & Roles

Cowboy CRM has 11 distinct roles, each with a precisely scoped permission set. Your role determines what you see, what you can do, and how the system routes leads and tasks to you.

All 11 System Roles

Roles are assigned by Admin at account creation and stored as role_id on the user record. Each role maps to a unique permission bitmask that controls module visibility, write access, and routing participation.

IDRole NameScope Summary
1AdminFull system access. Configures all modules, manages users, views all data, controls integrations and security settings.
2OwnerBusiness owner view. Access to all P&L data, broker performance, cross-floor analytics, and system-wide configuration.
3VP SalesSenior management. Full client list, all agent performance, finance summaries, broker data. Can create managers and team leaders.
4Team LeaderManages a specific agent group. Can view and reassign their team's leads, approve callbacks, create tasks, send team broadcast.
5Operations ManagerFloor operations view. Shift management, agent scheduling, workload balancing across campaigns. No finance write access.
6Chat RepChat-channel specialist. Handles client communications via Chat module. Limited client profile access — cannot see deposit data.
7AgentFront-line sales agent. Sees their assigned leads, tasks, shift schedule, personal gamification scores, and floor chat.
8SupportRetention and support specialist. Can access client profiles of existing depositors. Creates follow-up tasks and escalations.
9Content WriterEmail and communication template author. Access to Communications module for template creation. No client data access.
10Office ManagerOperational administration. User management, shift publishing, document uploads. No finance data access.
11Instructions ManagerKnowledge and compliance content. Manages the Documents module and training materials. Read-only on client data.
Roles are permanent at account creationA user's role cannot be changed via the UI after creation. Role changes require admin-level database access to modify the role_id field and a full session logout/re-login to take effect.

Routing Participation by Role

Not all roles participate in AI lead routing. The routing engine only considers users whose role is in the eligible set. Roles outside the routing pool are excluded automatically — no configuration needed.

RoleReceives AI-routed leads?Participates in Gamification?Chat access?
Admin (1)
Owner (2)
VP Sales (3)View only
Team Leader (4)Optional
Operations Manager (5)
Chat Rep (6)Chat channel only
Agent (7)✓ Primary
Support (8)✓ Depositors only
Content Writer (9)
Office Manager (10)
Instructions Manager (11)

Permission Matrix

Feature / ActionAgent (7)TL (4)VP Sales (3)Owner (2)Admin (1)
View own leads
View all leadsTeam only
Edit client recordsOwn onlyTeam only
View statistics / reportsOwn onlyTeam only
Manage shiftsLimited
View broker data
Finance / deposit dataView only
Manage usersLimited
System settings
Security audit logView
Chat & messaging
Gamification leaderboardViewView
Play embedded games
Module 01

Clients

The central client intelligence layer. 168,000+ client records with full interaction history, AI routing log, deposits, notes, and assignment history. This is where every lead lands and every deal is tracked.

168k+ records
Advanced filtering
AI-routed assignments

The Client List Screen

The main view shows all clients accessible to your role. Each row is one client. Click any row to open the full client profile. The AI routing log shows every decision the system made for this client.

cowyboymanager.com/clients1
+ New Client
⬇ Export
🔍 Filter
📅 Date Range
168,247 clients · Page 1 of 6,730
NameCountryStatusAgentAI ScoreLast Action
Klaus Weber🇩🇪 GermanyDepositedSarah M.942 min ago
Carlos Ruiz🇪🇸 SpainCallbackDaniel R.8114 min ago
Marie Dupont🇫🇷 FranceIn CallAnna K.76Now

Button Reference

+ New Client
Add a new client manually

Opens the new client form. AI will score and assign the client immediately on save.

⬇ Export
Export client list

Exports the current filtered view as CSV or Excel. Logged in the security audit trail.

🔍 Filter
Advanced filtering panel

Filter by country, status, agent, broker, campaign, date range, AI score range, deposit status.

📅 Date Range
Date range picker

Filter clients by registration date, last action date, or deposit date. Supports presets: Today, This Week, This Month.

Client Profile Fields

first_name / last_name
Text
Client's full name. Required on creation.
email
Email
Primary contact email. Used for communications dispatch.
phone
Phone
Primary phone. International format required. Used for call scheduling.
country / language
Select
Used by AI for geo-routing and language-based agent matching.
status
Select
Client's current operational status in the pipeline. Drives routing eligibility and task creation. See Deposit Status Pipeline for the 22 deposit-specific statuses.
tag
Enum (14 types)
Client classification tag: NA (No Answer) · NI (Not Interested) · CA (Call Again) · TR (Transfer) · RF (Referred) · LD (Lead) · EN (Enrolled) · WT (Waiting) · NAC (No Active Contact) · BTGP (Back to General Pool) · DUP (Duplicate) · MERGE (Merged record) · TECH (Technical issue) · DP (Deposited). Multiple tags can be applied; DP tag triggers the deposit pipeline.
priority
Enum
Lead handling priority: HIGH · NORMAL · LOW. Set by routing engine on arrival; overridable by manager. HIGH priority leads are surfaced first in agent task lists regardless of assignment order.
assigned_agent
Auto / Manual
Set by AI on creation. Can be manually overridden by manager or admin.
ai_score
Integer 0-100
Composite lead quality score calculated by the AI engine. Updates as interactions occur.
broker_id
Reference
Links client to the broker that sent or is managing this lead.
campaign_source
Text
Traffic source tag. Used for affiliate attribution and campaign ROI calculation.
notes
Text area
Free-text notes visible to assigned agent and managers. Timestamped entries appended.

Common Workflows

How to add a new lead manually

1
Click "+ New Client"

The new client drawer slides in from the right side of the screen.

2
Fill required fields

Name, phone, email, country, and language are required. Campaign source is optional but recommended for attribution.

3
Save client

On save, the AI engine scores the lead immediately and assigns it to the optimal agent currently on shift.

4
System triggers action

A task is created for the assigned agent. If a call window is active, the call is scheduled immediately. If not, it queues for the next optimal window.

Do not duplicateThe system checks for duplicate phone numbers on save. If a duplicate is detected, you will be shown the existing client record instead of creating a new entry.

Advanced Filtering System

The filter panel contains 40+ parameters organized across four tabs. Active filters appear as dismissible pills above the client list. Filter combinations can be saved as named presets and shared with other managers — eliminating the need to rebuild complex filter sets repeatedly.

Filter Tabs & Parameters

TabAvailable Filters
BasicCountry, status, language, assigned agent, broker
AdvancedDeposit status (FTD / no deposit / multiple deposits), FTD count range, campaign, traffic source, registration method
AI FiltersAI score range (e.g., 70–100), routing history (routed X+ times), predicted value tier (High / Medium / Low), last routing outcome
Date FiltersRegistration date, last action date, deposit date, callback scheduled date, first contact date. All support: exact range, "before", "after", and relative presets (Today / This Week / Last 30 Days)
Save filter presetsClick "Save as Preset" after building a filter combination. Give it a name (e.g., "German high-score no deposit last 7 days") and share it with other managers so they can apply it with a single click.

Bulk Actions

Select multiple clients using the row checkboxes, then use the bulk action bar that appears at the bottom of the screen. Bulk actions are role-restricted and logged in the audit trail with the full list of affected client IDs.

Bulk Reassign
Reassign to different agent

Select target agent from dropdown. AI routing weight is preserved — next auto-route event will re-evaluate.

Bulk Status Update
Change status across selection

Set all selected clients to a new status (e.g., bulk mark as "Callback" after a campaign event).

Bulk Tag
Apply campaign tag

Append a campaign or custom tag to all selected clients. Useful after manual import or batch reassignment.

Bulk Export
Export selected records

Exports only the selected clients as CSV. Export is logged with the client ID list and performing user's name.

Add to Sequence
Enroll in communication sequence

Adds all selected clients to a Communications sequence. System will send templated messages on the configured schedule.

Client Timeline View

Every client profile includes a chronological interaction timeline — a full audit trail of everything that happened to and with this client inside Cowboy CRM. The timeline is read-only and cannot be edited or deleted by any user.

Assignment event
Auto
When the client was assigned to an agent — by AI or manually — with the reason and agent name.
Status change
Auto
Every status transition: who changed it, from what, to what, and when.
Call attempt
Auto / Agent
Each logged call: outcome (answered / no answer / voicemail), duration, agent, timestamp.
Note added
Agent
Free-text note with author and timestamp. Agents see their own notes; managers see all.
Deposit event
Auto
Deposit amount (visible to manager/admin only), method, and attribution data at moment of deposit.
Email / SMS sent
Auto
Each communication dispatched: channel, template used, delivery status, open/click result.
Task created / completed
Auto
Task creation, deadline, completion or escalation events. Outcome note included if agent wrote one.
Agent changed
Auto
Every agent reassignment — includes previous agent, new agent, who triggered the change (AI or user), and timestamp.

Agents see only the timeline events that occurred during their assignment window. Managers and admins see the full history of the client across all agents and time periods.

Not Interested — 21 Real Reasons

When a client is tagged NI (Not Interested), the system requires a reason code to be selected. This is not cosmetic — NI reason data feeds campaign quality analysis and helps identify which traffic sources or brokers are generating low-intent leads. These are the 21 real reason codes in the system:

Already trading
Already with competitor
Bad timing
Budget constraints
Call dropped
Cannot afford minimum
Complained about calls
Different language
Do not call request
Elderly / vulnerable
Family objection
Fake lead / bot
Fear of losing money
Legal restrictions
Lost interest after research
No disposable income
Not the right product
Personal reasons
Previous bad experience
Scam concern / distrust
Wrong number / person
NI reason feeds campaign intelligenceAggregated NI reason data in the Reports module shows exactly why leads from specific brokers or campaigns are rejecting. "Fake lead / bot" spikes from a broker trigger a review flag. "Do not call request" records feed the DNC list automatically.

Transfer Types — 6 Real Modes

When a client is transferred between agents, departments, or to a broker, a transfer type must be selected. This categorizes the handoff and determines the receiving queue and notification logic. Six transfer types exist in the system:

Transfer TypeWhen It's UsedRouting Behavior
Live TransferAgent has client live on the phone and transfers immediately to another agent or specialistImmediate — recipient agent is notified and phone handoff occurs in real time
Online TransferClient is transferred via the online/chat channel rather than phoneChat session handed to the receiving agent. No phone call involved.
Support TransferClient needs to move from sales to the support/retention teamRe-assigned to a Support (role 8) agent. Sales agent's assignment ends.
Schedule call TransferTransfer + callback scheduled. Receiving agent gets a task with the agreed call timeTask auto-created for receiving agent with the scheduled callback time.
Email TransferTransfer notification and context sent to the receiving agent via email rather than in-system notificationUseful when the receiving agent is external or uses email as primary tool.
Try to callPassive transfer — receiving agent is asked to attempt contact rather than scheduled for a definite callLow-priority task created. No time commitment. Receiving agent works it when available.

AI Routing Log

The AI routing log is a sub-panel on the client profile that shows every routing decision the system made for this client — including decisions the agent never sees. This is the most powerful diagnostic tool for understanding why a client ended up with a particular agent.

ColumnWhat It Shows
TimestampExact datetime of the routing decision
Agents consideredAll agents evaluated in that routing cycle and their eligibility status (on shift / language match / campaign match)
Scores assignedThe composite score each eligible agent had at that moment — showing exactly why one agent ranked above another
Final assignmentThe agent selected and the reason: highest score / language priority / campaign priority / manual override
Override flagWhether a manager manually overrode the AI recommendation, and which manager made the change
Module 05

Shifts

The autonomous shift management layer. Cowboy CRM builds and manages shifts automatically based on historical conversion data. Managers review and can override — but rarely need to.

How Shift Intelligence Works

Cowboy CRM analyzes years of call center data to identify which hours, days, and shift configurations produce the highest FTD rates for each campaign. It then builds shifts accordingly and assigns the best-matched agents based on language, performance score, and availability.

Fully autonomous by defaultUnless a manager creates a shift manually, Cowboy CRM generates all shifts for the next 7 days every Sunday night. Manual creation is available but rarely needed.

Shift Screen Buttons

+ New Shift
Create shift manually

Override AI generation. Set date, time window, campaign, and agents. Use for special events or emergencies.

📊 Heatmap
Conversion heatmap

Shows conversion rates by hour for the selected date range. Used to understand optimal shift windows.

Reassign
Temporary reassignment

Move an agent from one shift to another for today only. System keeps permanent assignment unchanged.

Mark Absent
Mark agent absent

Removes agent from active shift. System auto-redistributes their assigned leads to other agents on shift.

Shift Fields

shift_name
Text
Label for the shift. Auto-generated as "Morning Shift — Mon Mar 31" or custom name.
start_time / end_time
Time
Shift window. AI schedules calls only during active shift windows for assigned agents.
assigned_agents
Multi-select
Agents on this shift. AI populates based on performance and language. Managers can override.
campaign
Reference
Campaign this shift serves. Determines language requirements and routing rules for the window.
conversion_target
Percentage
Target FTD rate for this shift. Used to flag underperformance and trigger manager alerts.

AI Shift Generation Logic

Every Sunday at 23:00, Cowboy CRM autonomously builds shifts for the coming week. The algorithm evaluates historical data and current resource availability to generate an optimized schedule that no human planner could match in speed or granularity.

Inputs to the Algorithm

Historical FTD rates
Data input
Conversion rates broken down by hour of day and day of week for each campaign over the trailing 90 days.
Agent availability
Data input
Active agents, their language capabilities, and their campaign assignments. Agents on leave are excluded automatically.
Target coverage ratios
Setting
Minimum agents per language per campaign per shift window. Configured in Settings → Shift Rules.
Campaign calling windows
Setting
Allowed calling hours per campaign per country — respects local regulations and time zone boundaries.

The generated schedule is delivered to the manager inbox at 07:00 Monday as an email summary listing all shifts for the week, the agents assigned to each, conversion targets, and any coverage gaps the AI could not fill due to insufficient agent availability.

Fully autonomous by defaultUnless a manager creates a shift manually, Cowboy CRM generates all shifts for the next 7 days every Sunday night. Manual creation is available but rarely needed.

Conversion Heatmap

The heatmap is a 24-hour × 7-day grid showing historical conversion rates for the selected date range. It is the primary visual tool managers use to understand optimal shift windows and validate the AI's schedule decisions.

ColorMeaningAction
GreenHigh conversion rate for this hour/day cellPrioritize agent coverage during this window
AmberMedium conversion rateStandard coverage appropriate
RedLow conversion rateMinimal agent coverage; shift resources to green windows
GreyNo data (new campaign or no calls in this window)Defer to campaign calling window rules until data accumulates

Hover over any cell in the heatmap to see the exact FTD count and conversion rate for that hour/day combination. Use this detail view to manually refine shift start and end times when the AI-generated windows need fine-tuning.

Absence & Coverage Management

When an agent is marked absent mid-shift, Cowboy CRM immediately redistributes their assigned leads and triggers coverage alerts if the shift drops below the configured minimum agent count.

1
Mark agent absent

Manager opens Shifts module, finds the agent on the active shift, and clicks "Mark Absent." Takes effect immediately.

2
Leads auto-redistributed

System first tries agents on the same shift with the same campaign and language. Fallback: any on-shift agent with matching language. Fallback-of-last-resort: any on-shift agent on that campaign.

3
Coverage threshold check

If the shift drops below the minimum agent threshold after the absence, a manager notification is triggered immediately.

4
Absence logged in agent profile

The absence event is stored in the agent's shift history. Unexcused absences reduce the agent's punctuality score for the current period.

Shift Performance Review

After each shift window closes, the system automatically generates a shift performance summary and sends it to the managing team. These summaries accumulate in the Shift History panel for trend analysis.

FTDs in window
Count
Total first-time deposits recorded during this shift's time window.
Agents on shift
Count
How many agents were active during the shift vs. how many were assigned (coverage efficiency).
Top performer
Agent name
Agent with the highest FTD count or conversion rate during this specific shift window.
Coverage %
Percentage
Actual agent-hours worked / planned agent-hours. Below 80% triggers a review flag.
Conversion rate vs. target
Delta
Actual FTD rate compared to the target set when the shift was generated. Green if above, red if below.
Module 09

Games & Gamification

The full gamification engine. Live leaderboards, achievement badges, personal missions, and team competitions — all connected to real performance data and updated in real time via Socket.IO.

The Live Leaderboard

The main Games screen shows the live leaderboard — all active agents ranked by composite score. The leaderboard updates in real time via Socket.IO: no page refresh needed. Every call logged, task completed, quiz submitted, or shift clock-in triggers an immediate score recalculation and re-rank.

Leaderboard Columns

ColumnWhat It ShowsNotes
RankCurrent position on the floorUpdates live as scores change
Agent NameAgent display name and avatar initialAgents see their own row highlighted
ScoreComposite 0–100 score for the current periodCalculated from all 5 components
ChangeUp/down arrow + delta since yesterday's scoreGreen arrow = improving, red = declining
LevelBronze / Silver / Gold / Platinum / Elite tier badgeAffects lead priority assignment

The time period selector — Today / This Week / This Month / All Time — is available at the top of the leaderboard. Agents see their own rank and how many points separate them from the next position above. Managers see the full floor ranking.

Score updates live — no refresh neededEvery time an agent logs a call, completes a task, takes a quiz, or clocks in on time, their score updates immediately and the leaderboard re-ranks in real time via the Viki Socket.IO engine.

Score Calculation — Detailed

The composite score is a weighted sum of five independent components. Each component has its own measurement model and reset cadence.

ComponentWeightMeasurementScoring Detail
Conversions / FTDs40%FTD rate vs. total assigned leads in periodAgent with 20% FTD rate scores higher than one with 8%. Resets each period start.
Response time20%Minutes from lead assignment to first contact attemptUnder 5 min = 100pts · 5–15 min = 75pts · 15–30 min = 50pts · 30+ min = 25pts
Quiz scores15%Running average of all quizzes completed in periodEach quiz result averaged in. Missing a quiz counts as zero for that quiz.
Punctuality15%On-time clock-ins / total shift starts in periodClock-in within the configured grace window (default 5 min) counts as on time.
Callback compliance10%Callbacks completed on time vs. total scheduledCallback completed within ±15 minutes of agreed time counts. Missed = zero credit.

The 6 Playable Games (Real)

Beyond the performance leaderboard, Cowboy CRM includes 6 fully playable arcade games accessible during breaks. These are real embedded games — not simulations. Each game requires a backend permission check before it can be opened. Games are tab-based inside the Gamification module.

Tab AliasGame NamePermission KeyNotes
TETRISTetrisusers/games/perm/checkClassic block-stacking game. Embedded full implementation.
BREAKERBrick Breakerusers/games/perm/checkBall-and-paddle breakout-style game.
PACMANPacmanusers/games/perm/checkFull maze-navigation Pacman clone.
SNAKESnakeusers/games/perm/checkClassic snake growth game.
TICTic Tac Toeusers/games/perm/checkTwo-player or vs-AI tic tac toe.
PONGPongusers/games/perm/checkClassic two-paddle Pong game.
Permission-gated accessEvery game tab calls users/games/perm/check on load. Admin can enable or disable individual games per user role in Settings → Gamification. Games inaccessible to a role are hidden, not just disabled.

Agent Tiers

Every agent is assigned to one of 4 tiers based on their composite performance score. Tiers are recalculated in real time as scores change and directly affect which leads the agent receives — Tier1 agents get the highest-value leads first.

TierStandingLead Priority Effect
Tier1Top performersHighest-value leads always route to Tier1 agents first. First pick on every incoming batch.
Tier2Strong performersPriority routing. High-AI-score leads prefer Tier2+ agents when Tier1 agents are at capacity.
Tier3Average performersStandard priority. Receives leads ahead of Tier4 agents on the same campaign.
Tier4New or low performersStandard lead pool. Lead flow is sufficient to work from, but top-tier leads are withheld until score improves.

Badges & Achievements

Badges are awarded automatically when milestones are reached. They appear on the agent's profile, on the leaderboard row, and in the floor activity feed. Each badge has a rarity level that reflects how difficult it is to earn.

🥇
First Deposit
First FTD of the month
Common
🔥
Hot Streak
5 conversions in a row
Rare
Fast Responder
Response under 60 seconds
Rare
🎯
Perfect Week
Zero missed callbacks for 7 days
Rare
📝
Quiz Master
100% on a compliance quiz
Common
Always On Time
30 days perfect punctuality
Epic
💯
Century Club
100+ FTDs in a single month
Epic
🌍
Language Master
Converted leads in 3+ languages
Rare
🌙
Night Owl
Most conversions during night shift
Common
🏎️
Speed Demon
Fastest avg response time this week
Common
📈
Comeback King
Biggest score improvement week-over-week
Rare
🤝
Team Player
Most shared client cards in Chat this week
Common

Personal Missions

Each morning, the system assigns a fresh set of daily missions to every active agent. Missions are generated based on the agent's historical performance baseline — agents with lower callback compliance get callback missions, agents with slow response get response-time missions. Completing a mission triggers an instant in-system notification and score bonus.

Example Daily Missions

Contact push
Auto
"Contact 15 new leads today" — scales to agent's historical daily contact average.
Callback target
Auto
"Achieve 3 callbacks before noon" — time-boxed target to drive morning activity.
Quiz completion
Auto
"Complete today's compliance quiz" — always assigned if a new quiz was published.
Response standard
Auto
"Maintain sub-10-minute average response today" — only assigned to agents with slow response history.
Custom mission
Manual
Manager can assign any custom mission to a specific agent with a custom title, description, and deadline.

Team Competitions

Managers can create structured competitions between agent groups. A competition has a clear metric, duration, and optional prize. The system handles all scoring, ranking, and winner declaration automatically.

Competition Setup Fields

name
Text
Competition title shown to all participants on the leaderboard.
start / end date
DateTime
Competition window. Scores calculated from events within this window only.
metric
Enum
What is being competed on: FTDs / Response time / Conversion rate. One metric per competition.
teams
Multi-select
Group agents by: same campaign, manually selected agents, or shift-based groups.
prize_description
Text
Optional prize text displayed to agents during the competition for motivation.

A live team leaderboard is visible to all participants during the competition window. Competition history — all past competitions with final rankings and winners — is stored in the Games module for reference.

Floor Activity Feed

The activity feed is a live event ticker in the Games/Chat panel. It creates a real-time competitive atmosphere on the floor by broadcasting significant events to everyone logged in.

Badge earned
Auto-post
Example: "Sarah M. just earned Hot Streak badge 🔥" — visible to all agents on the floor.
FTD conversion
Auto-post
Example: "Klaus Weber deposited — Daniel R. gets credit" — celebrates the conversion with full attribution.
Competition milestone
Auto-post
Example: "Team Alpha is now #1 in the weekly competition" — updates when rank changes occur.
Mission complete
Auto-post
Agent completes a daily mission — posted to the floor feed as a morale signal.

Manager can mute specific event types from the feed in Settings → Gamification. This is useful for quieter call floors or compliance-sensitive environments where broadcast announcements are inappropriate during call hours.

Manager Tools

Bonus Period
Set score multiplier window

Define a time window (e.g., "This Friday 2–5pm") where all FTDs earn double gamification points. Agents see the countdown timer on the leaderboard.

Announce Winner
Broadcast winner to floor

One click sends a floor broadcast with competition winner name and metric result. Triggers a badge award if configured.

Score Breakdown
View individual agent components

Opens a detailed breakdown of any agent's composite score: each of the 5 components with individual sub-scores and trend.

Reset Period
Start new competition month

Resets all period scores to zero for a clean competition slate. Previous period scores are archived and visible in history.

Export Leaderboard
PDF for display or meeting

Generates a formatted PDF of the current leaderboard — suitable for printing on the office wall or presenting in a team meeting.

Module 10

Reports

Automated reporting across every dimension of your operation. Scheduled FTD reports, campaign P&L exports, broker performance digests, agent conversion trends, and executive summaries — generated automatically and delivered on schedule.

Automated daily & weekly delivery
Scheduled exports to email or SFTP
All dimensions: agent, campaign, broker, shift

Dashboard Panels

PanelShowsWho Sees It
FTD CounterLive count of first-time deposits today, this week, this monthManager, Admin, Viewer
Agent PerformanceConversion rates, call counts, response times per agentManager, Admin
Broker ComparisonSide-by-side FTD rate and revenue per brokerManager, Admin
Conversion HeatmapHour-by-hour conversion rate for current weekManager, Admin
Campaign ROIRevenue vs. cost per campaign, by broker and channelAdmin, Viewer
My Stats (Agent)Personal score, rank, conversions, callbacks pendingAgent
Reports are automaticDaily and weekly summary reports are emailed to managers and admins automatically. No manual export needed. Reports include: FTD count, top agents, broker performance, and shift conversion breakdown.

Scheduled Report Types

ReportFrequencyRecipientsFormat
FTD Daily SummaryDaily at end of shiftManager, AdminEmail + in-system
Agent Performance WeeklyEvery Monday 08:00Manager, AdminEmail with CSV attachment
Broker Performance DigestDailyAdminEmail
Campaign P&L Monthly1st of each monthAdmin, ViewerEmail with Excel attachment
Shift Conversion BreakdownAfter each shift closesManager, AdminIn-system + optional email
Security Audit WeeklyEvery MondayAdminEmail

On-Demand Reports

In addition to scheduled reports, all report types can be generated on-demand from the Reports module. Apply any filter combination — date range, broker, campaign, agent, country — then export as CSV or Excel. All exports are logged in the security audit trail.

Module 06

Tasks

AI-assigned tasks with subtasks, timers, and priority scores. Every follow-up, callback, and compliance action is tracked, escalated, and logged automatically.

Auto-created on every lead assignment
Escalates overdue items to team leads
Priority scoring by lead value + urgency

How Tasks Are Created

Tasks in Cowboy CRM are not created manually by default — the system generates them automatically from system events. When a new lead is assigned to an agent, Cowboy CRM creates a task: "Initial contact — call within 2 hours." If the client requests a callback, a new task is created with the exact callback time. Overdue compliance actions are auto-escalated.

Zero manual setup requiredThe system creates tasks from lead events, call outcomes, compliance triggers, and manager assignments. Agents arrive to a pre-populated task list every shift.

Task Priority Scoring

FactorImpactDescription
Lead AI scoreHighLeads with higher AI scores generate higher-priority tasks
Time since assignmentHighPriority increases as time passes — first hour is critical
Campaign urgencyMediumSome campaigns flag all leads as urgent — overrides other scoring
Deposit potentialMediumClients with prior deposits or high-value profiles ranked higher
Callback commitmentCriticalTasks with a specific client-requested time are highest priority when due

Button Reference

+ New Task
Create task manually

Manager or admin can create a one-off task for any agent. Specify deadline, priority, and description.

Complete
Mark task complete

Logs outcome and closes task. Completion time is recorded for punctuality scoring.

Reschedule
Move task deadline

Extend or change the task deadline. Requires a reason note. Logged in audit trail.

Escalate
Escalate to manager

Flags the task for manager review. Used when a client situation requires senior intervention.

Task Fields

title
Text
Task description. Auto-generated or manually entered.
assigned_to
Reference
Agent responsible. Set by AI on creation, overridable by manager.
client_id
Reference
The client this task relates to. Clicking opens the client profile.
deadline
DateTime
When the task must be completed. Overdue tasks trigger escalation chain.
priority
Enum
Task priority: HIGH · NORMAL · LOW. Set by the system based on lead value, urgency, and campaign configuration. Overridable by manager.
sub_tasks
JSON array
Optional checklist of steps inside the task. Each subtask has its own completion state.
timer_running
Boolean
Active timer tracking how long the agent has been working on this task.
outcome_note
Text
Required on completion. Agent must describe the outcome before marking complete.

Task Sub-Types — 12 Real Types

Every task has a sub-type that determines its behavior, escalation rules, and how it appears in the agent's task list. The system has 12 real sub-types:

Sub-TypeCategoryDescription
TaskStandardGeneral work item. Manual or auto-created. Flexible deadline and priority settings.
MeetingCalendarScheduled meeting with a specific time slot. Syncs with the Calendar module.
OutlinePlanningHigh-level planning item. Typically used by managers and team leaders for campaign or project planning.
AnnouncementBroadcastFloor-wide announcement item. Creates a notification visible to all agents in scope.
Open ShiftShiftsShift coverage task. Created when a shift slot is open and needs to be filled. Linked to Shifts module.
QuestionSupportAgent or client question requiring an answer. Tracked until resolved with a documented response.
Incoming EmailCommunicationsAuto-created when a client email is received. Links to the email thread in Communications module.
Outgoing EmailCommunicationsTracks an outgoing email that requires follow-up confirmation or a reply check.
Draft TaskDraftTask saved as draft — not yet active. Does not trigger notifications or affect agent task count.
Draft EmailDraftEmail draft saved in the task system for completion later. Auto-linked to client profile.
Draft MeetingDraftMeeting placeholder saved before final confirmation of time and participants.
EventCalendarScheduled event with no required action. Visible on Calendar module for awareness and planning.

Task Repeat Types — 9 Schedules

Recurring tasks can be set to any of 9 repeat patterns. The system creates new task instances automatically according to the selected schedule:

CodeScheduleUse Case
ONCEOne time onlyStandard non-recurring task. Default for most auto-created tasks.
MON-FRIEvery weekdayDaily recurring tasks that follow the business week. Pauses on weekends automatically.
WWeeklyRepeats every 7 days from the creation date. Used for weekly check-in calls.
MMonthlyRepeats on the same day of each month. Common for monthly client review tasks.
QQuarterlyEvery 3 months. Used for compliance review cycles and quarterly performance tasks.
H-YHalf-yearlyEvery 6 months. Used for bi-annual client account reviews.
E-YEvery yearAnnual recurring tasks. Used for anniversary contacts and yearly compliance updates.
ANAnniversaryTriggers on the same calendar date each year. Used for client anniversary contact campaigns.
OP-SHFTOpen shiftLinked to the Shifts module. Recurs whenever an open shift slot is published in the agent's schedule.

Task Templates

Admin can define reusable task templates for common multi-step workflows. Applying a template to a client creates the parent task and all defined subtasks in a single action — eliminating repetitive manual setup for routine processes.

Example Built-In Templates

Template NameSubtasks IncludedDefault Deadline Window
Initial contact sequence1st call attempt · 2nd call attempt · Send intro email · Log outcome48 hours
Deposit follow-up chainConfirmation call · Verify receipt · Send deposit confirmation · Schedule next call24 hours
Compliance reviewReview KYC docs · Verify identity · Update compliance status · Manager sign-off72 hours (escalates faster)
Re-engagement sequenceReview last interaction · Personalized callback · Send re-engagement email · Update status7 days
template_title
Text
Template name shown in the "Apply Template" picker on the client profile and task creation form.
subtask_list
JSON array
Ordered list of subtask titles. Each becomes a subtask on the created task, preserving the defined order.
default_deadline
Duration
Hours from task creation to deadline. Applied automatically when template is used; overridable at creation time.
default_priority
1–5 integer
Starting priority for tasks created from this template. AI re-scores based on client AI score immediately after creation.

Escalation Chain

Overdue tasks automatically escalate through a four-level chain. Each level notifies a progressively senior person. Compliance tasks have an accelerated chain — any overdue compliance task goes directly to manager notification regardless of how long it has been overdue.

1
Level 1 — Overdue 1 hour

Notification sent to the assigned agent. Task highlighted in red in their task list. Score impact begins after this threshold.

2
Level 2 — Overdue 4 hours

Notification sent to the agent's team lead (the manager assigned to their campaign). Task flagged in team overview.

3
Level 3 — Overdue 24 hours

Notification sent to the manager (Admin or Manager role). Task flagged in the agent's permanent profile as a missed SLA event. Affects agent performance score.

4
Level 4 — Compliance tasks (any overdue)

Compliance-tagged tasks skip Levels 1–3 and trigger immediate manager notification. Client is flagged for compliance review. Cannot be dismissed without a manager acknowledge action.

Escalation chain is configurableHour thresholds for Levels 1, 2, and 3 can be adjusted in Settings → Task Rules. Different thresholds can be set per task type (standard vs. compliance vs. callback).

Subtask System

Tasks can contain up to 20 subtasks. Each subtask is a discrete step with its own completion state. The parent task cannot be marked complete until all required subtasks are finished — preventing agents from closing tasks with incomplete steps.

subtask_title
Text
Description of the step. Shown as a checkbox item in the task detail view.
completed
Boolean
Checked by the agent when the step is done. Completion time is logged individually in the audit trail.
assignee
Optional reference
Subtask can be assigned to a different agent than the parent task — useful for hand-off workflows or compliance sign-off steps.
required
Boolean
Required subtasks block parent task completion. Optional subtasks can be skipped with a reason note.

Each subtask completion is logged individually in the audit trail — timestamp, who completed it, and any note attached. This granular logging is particularly valuable for compliance workflows where regulators may request proof that each step was completed independently and in sequence.

Module 07

Users

Agent profiles, role permissions, performance history, language capabilities, and composite scoring. Every user in the system is a data model — not just a name in a list.

11 roles: Admin, Owner, VP Sales, TL, Ops Mgr, Chat Rep, Agent, Support, Content Writer, Office Mgr, Instructions Mgr
username_special login (not email)
Composite score drives routing decisions

The Agent Profile

Each user account has a full performance profile built automatically from every interaction. The composite score — visible on the agent's profile and used in routing — is calculated from conversion rate, response time, quiz scores, punctuality, and callback compliance.

Routing uses this data in real timeAgent composite scores are recalculated continuously. A score drop from missed callbacks or failed quizzes immediately affects which leads the agent receives — even mid-shift.

Creating a New User

1
Navigate to Users module

Admin role required. Click "+ New User" in the top right of the Users screen.

2
Set username_special

This is the login credential — NOT an email. Choose a format like firstname_lastname or an ID. Cannot be changed after creation without admin database access.

3
Assign role and languages

Select role from the 11 available roles (Agent, Team Leader, VP Sales, Admin, etc.). Set language capabilities — the routing engine uses these for lead assignment. Roles 7 and 8 (Agent, Support) are the primary routing-eligible roles.

4
Assign to campaigns

Select which campaigns this agent works. Leads from unassigned campaigns will not route to this agent.

5
Share credentials

Provide the username_special and initial password to the agent. They login at cowyboymanager.com — no email setup needed.

User Fields

username_special
String
The login credential. Not an email. Unique per system. Used in the enc_user encryption chain.
role
Enum (1–11)
Admin(1) / Owner(2) / VP Sales(3) / Team Leader(4) / Operations Manager(5) / Chat Rep(6) / Agent(7) / Support(8) / Content Writer(9) / Office Manager(10) / Instructions Manager(11). Controls module visibility, write access, and routing participation.
languages
Multi-select
Languages the agent can handle. AI only routes leads in languages the agent has enabled.
campaigns
Multi-select
Campaign assignments. Restricts lead routing to specific campaigns for this agent.
composite_score
Float 0-100
Auto-calculated. Drives routing priority. Displayed on leaderboard. Updates continuously.
active
Boolean
Inactive users cannot log in and are excluded from routing. Deactivate instead of deleting to preserve history.

Agent Performance Profile

Every agent has a comprehensive performance profile accessible to managers and admins. This is the most detailed view of any individual agent in the system — used for coaching, routing decisions, commission calculation, and HR reviews.

Performance Profile Sections

SectionWhat It ShowsWho Can See
Current period score breakdownAll 5 score components with individual sub-scores and their weight contribution to the compositeManager, Admin
Historical score trend30-day sparkline of composite score — shows whether agent is improving or decliningManager, Admin
FTD history by monthMonthly FTD count for the trailing 12 months with visual bar chartManager, Admin
Language distributionBreakdown of converted leads by language — shows which languages drive the agent's conversionsManager, Admin
Campaign performanceConversion rate and FTD count per campaign this agent worksManager, Admin
Floor comparisonAgent's score vs. floor average across all 5 components — highlights specific strengths and weaknessesManager, Admin

The performance profile is also the input to the AI routing engine. The composite score, language capabilities, and campaign assignments are all read from this profile every time a routing decision is made. Keeping the profile accurate — especially language levels and campaign assignments — directly impacts the quality of the agent's lead flow.

Routing uses this data in real timeAgent composite scores are recalculated continuously. A score drop from missed callbacks or failed quizzes immediately affects which leads the agent receives — even mid-shift.

Language Capabilities Management

Language assignment is one of the most critical settings on any agent profile. The routing engine uses language level — not just language presence — to decide whether to route a lead to an agent.

Language LevelRouting Behavior
NativeEligible for all leads in that language. Preferred over Fluent agents when AI scores are equal.
FluentEligible for all leads in that language. Receives leads when no Native-level agent is available.
BasicOnly eligible if no Native or Fluent agent is on shift. Used as a last-resort fallback to prevent a lead from going unworked.

Manager can record a language assessment result in the agent's profile — a free-text note plus a date — to document when the level was last evaluated. This assessment record is visible in the routing decision log, which shows exactly why a language-level determination was made for each routing event.

Review language levels regularlyAn agent incorrectly set to "Native" for a language they only have basic ability in will receive high-value leads they cannot convert effectively. Audit language levels quarterly or after any conversion rate drop in a specific language.

Bulk User Management

For operations with large agent rosters or frequent onboarding cycles, bulk tools eliminate the overhead of managing users one at a time.

Bulk Deactivate
Deactivate multiple users

Select agents with checkboxes and bulk-deactivate. All sessions terminated, leads redistributed per the configured fallback. Admin role required.

Bulk Campaign Assign
Assign to campaigns at scale

Select multiple agents and assign them to a campaign in one action. Useful when launching a new campaign with an existing team.

Import from CSV
Bulk user creation via CSV

Upload a CSV with columns: name, username_special, role, languages. System creates accounts and sends credentials. Invalid rows flagged in an error report.

Export User List
HR export with performance data

Exports all users with: name, role, languages, composite score, FTD count, commission YTD. Suitable for HR review and payroll systems.

Module 02

Finance

Deposit tracking, FTD attribution, campaign P&L, and multi-currency reporting. Every dollar is attributed to its source across five dimensions simultaneously.

5-dimensional attribution per deposit
Multi-currency with live FX conversion
Anomaly detection on unusual patterns

Deposit Entry & Recording

Deposits enter the system through three paths: manual entry by an authorized agent or manager, automatic push from a broker API (broker sends a POST on payment confirmation), or agent-logged deposit after a verbal confirmation call. All three paths route through the same validation and attribution pipeline.

Deposit Form Fields

amount
Decimal
Deposit value. Required. Stored in original currency; converted to base currency for reports.
currency_id
Select
Currency of the deposit. Supported currencies: USD, EUR, GBP, ILS. Stored as currency_id linking to the currencies table. Converted to base currency for reports.
user_id
Reference
Links deposit to the client record (called user_id internally). Required. Triggers the 5-dimensional attribution chain on save.
method_id
Enum (17 types)
Real payment methods in the system: Credit Card · Skrill · Neteller · EFT Pay · Sofort · Wire Transfer · Other payment · POLI · Ideal · Cash · Deductible · PayPal Balance · Loan · Mix · Balance · Deferred · Check.
sec_user_id
Reference
Secondary user (broker-side contact). Links deposit to the broker relationship chain for CPA and performance tracking.
campaign_id
Reference
Campaign this deposit is attributed to. Links into the Brokers module for broker-campaign P&L calculation.
purchase_date
DateTime
The actual date of the deposit event. Separate from created_at — important for retroactive logging.
purchase_details
Text
Free-text detail about the deposit. Agent notes, confirmation codes, unusual circumstances.
purchase_type_id
Enum
Classification of the deposit type (FTD, re-deposit, bonus, etc.) used for P&L segmentation.
reason_id
Reference
Reason category for this deposit event — used in statistical analysis and campaign comparison reports.
log
JSON
System-generated audit log attached to the record. Captures every change, who made it, and when. Immutable after write.
Finance module is Admin/Manager onlyAgents cannot see deposit amounts, campaign P&L, or broker revenue data. This restriction is enforced at the API level — not just the UI.

5-Dimensional Attribution

When a deposit is saved, Cowboy CRM attributes it across five dimensions simultaneously. This is not cosmetic labeling — each attribution feeds a different downstream calculation engine. All five run at the moment of save, not in a nightly batch.

DimensionHow It's CapturedDownstream Use
AgentThe agent currently assigned to that client at deposit timeCommission calculation, performance scoring, routing weight update, agent P&L
CampaignCampaign tag stored on the client record at creationCampaign ROI, budget vs. actual, CPL tracking, campaign comparison reports
Brokerbroker_id on the client record — set when lead was deliveredBroker performance score, auto-rebalancing trigger, CPA net calculation
Traffic channelcampaign_source tag on the client (affiliate ID + channel)Affiliate CPA calculation, postback/pixel firing, channel ROI comparison
Shift slotActive shift window at the time of deposit timestampShift P&L, optimal staffing analysis, shift ROI vs. cost of coverage
Attribution is write-onceOnce a deposit is attributed and saved, the dimension tags are locked. Retroactive attribution changes require admin database access and are logged in the audit trail.

Deposit Status Pipeline — All 22 States

Every deposit moves through a lifecycle tracked by its status. Cowboy CRM has 22 real deposit statuses — each representing a specific operational state in the deposit pipeline. These are the exact status codes from the production system:

StatusMeaning & When It's Set
CHECKDeposit is pending verification. Default state when a new deposit is first logged — awaiting confirmation from broker.
TRADEDClient has traded with the deposit. Funds moved into trading activity on the broker side.
DONE SALESSales agent has completed their role in this deposit. Handoff point from sales to support/retention.
REMOVEDDeposit was removed from active tracking. Client did not proceed or deposit was voided.
APPROVEDDeposit confirmed and approved by broker. Funds cleared and attributed. Commission calculation triggers on this status.
CC-DECLINECredit card payment declined by the processor. Follow-up action required — agent prompted to contact client about alternative payment.
DEPOSIT INDeposit received by broker. Confirmation pending — intermediate state between CHECK and APPROVED.
DONE SUPPORTSupport team has completed their retention action on this deposit. Client has been handled post-deposit.
BACK TO SALESSupport or broker returned this client to the sales queue. New agent assignment may follow.
SCHEDULED CALL SUPPORTSupport has scheduled a callback specifically about this deposit. Task auto-created with the scheduled time.
DOCUMENTSBroker or compliance requires documents from the client before the deposit can be processed. Document checklist triggered.
SCHEDULED CALL AGENTAssigned agent has a scheduled callback about this deposit. Time-based task created.
SCHEDULED CALL BROKERBroker has scheduled a call with the client. Deposit is in broker's hands temporarily.
SCHEDULED CALL BROKER REQUESTThe broker has requested a call be scheduled. CRM will coordinate timing.
CHECKING WITH A BROKERAwaiting response from broker on deposit status. Common when broker platform has delays.
REMOVED BY MISTAKEPreviously removed deposit was removed in error. Re-activated with a correction note in the log.
NOT COMPLETEDClient initiated the deposit process but did not complete payment. Follow-up sequence triggered.
WDWithdrawal request submitted by client. Deposit amount is being processed for return.
SAVED WDAgent successfully saved a withdrawal — convinced client to keep funds deposited. Counts as retention win.
IN-PROGRESSDeposit is actively being processed by the broker. Intermediate state — status will update to APPROVED or REMOVED.
DEPOSIT-FROM-BROKERDeposit originated directly from the broker side rather than through an agent's client interaction.
SUPPORT STARTEDSupport team has begun working this deposit. Status set when support agent opens the deposit record.
Commission calculation triggers on APPROVEDThe commission engine fires when a deposit status changes to APPROVED. Earlier statuses do not trigger commission — only when the broker confirms the funds are cleared and the deposit is approved.

Agent Commissions

Commission calculation runs automatically when a deposit reaches APPROVED status. The system looks up the applicable commission structure for that agent, calculates the owed amount, and appends it to the agent's commission ledger for the current period.

TypeHow It WorksExample
Percentage-basedFixed % of each deposit amount attributed to the agentCommission % is configurable per agent in their profile settings — no system-default percentage is imposed.
Flat per FTDFixed amount per first-time deposit regardless of sizeFlat amount per FTD configurable per agent. Useful for standardized commission structures across a team.
Tiered (threshold)Rate increases as FTD count reaches configured thresholds in the periodThresholds and rates fully configurable per agent. Admin sets the breakpoints and corresponding rates.

Commission reports are exportable per agent, per period, in CSV format — formatted for direct import into most payroll systems. Managers can view commission totals per agent in real time without waiting for end-of-period reports.

Chargeback Management

A chargeback is a reversed deposit — a client's payment is disputed and returned. In Cowboy CRM, chargebacks are logged as negative entries in the deposit ledger and are tracked per broker, per agent, and per campaign.

client
Reference
The client whose deposit was reversed. Client record is flagged with chargeback indicator.
amount
Decimal
Amount reversed. Deducted from net revenue calculations and from agent commission if commission was already credited.
chargeback_date
DateTime
Date the reversal was confirmed. Used for period-accurate P&L calculation.
reason
Text
Reason code or free-text. Common values: Fraud, Dispute, Incorrect charge, Unauthorized.
broker
Reference
Broker linked to the original deposit. Chargeback increments that broker's chargeback ratio immediately.

Admin can configure a chargeback ratio alert threshold per broker. When a broker exceeds the threshold (e.g., 5% chargeback rate on trailing 30 days), the system sends an admin notification and flags the broker for review. Net revenue in all P&L reports is always calculated as gross deposits minus chargebacks — no manual adjustment needed.

P&L Reporting

The P&L module gives a profit-and-loss breakdown across every operational dimension. All figures are net of chargebacks and updated in real time as deposits are logged.

P&L ViewRevenue SourceCost SourceKey Metric
By CampaignAll deposits attributed to this campaignAd spend + CPL entries logged in SettingsNet margin %, ROI
By BrokerAll deposits from this broker's leadsCPA fee agreed with brokerNet broker value, revenue per lead
By ShiftAll deposits during this shift windowStaffing cost (hours × rate) if configuredShift ROI, revenue per agent-hour
By AgentAll deposits attributed to this agentAgent compensation (salary + commission)Revenue vs. cost ratio

Period comparison is built into every P&L view: current month vs. last month, current week vs. last week. Color-coded deltas show whether each metric improved or declined. Exportable as Excel for board-level reporting.

Multi-Currency Support

Every deposit stores its original currency alongside its converted base-currency value. FX rates are configured in Settings → FX Rates by an admin. The four native supported currencies are: USD, EUR, GBP, ILS. All P&L reports display in base currency with the per-deposit original currency available in the drill-down view. Currency is stored as currency_id linking to the currencies table.

FX rates are static, not liveCowboy CRM does not pull live FX rates automatically. Admin sets and updates rates manually in Settings. For high-volume multi-currency operations, update rates at least weekly.

Anomaly Detection

The Finance anomaly engine monitors every incoming deposit against a set of behavioral rules. Flagged deposits appear in the Anomaly Flags panel and trigger an immediate admin notification. The deposit is not blocked — it is logged and flagged for review.

Unusual amount
Auto-flag
Deposit is 3× or more the average deposit amount for this client's profile or campaign segment.
Rapid-repeat deposit
Auto-flag
Same client makes multiple deposits within a short configurable window (default: 2+ deposits in 1 hour).
IP geography mismatch
Auto-flag
Deposit logged from an IP that does not match the client's registered country. Possible proxy or spoofed location.
Off-hours deposit
Log
Deposit recorded outside the campaign's configured calling window for that geography. Logged with elevated visibility.

Button Reference

New Deposit
Log a deposit manually

Opens deposit form. All 5 attribution dimensions are captured automatically based on the linked client record.

FTD Report
First Time Deposit report

Daily FTD count by agent, broker, campaign, and time of day. Exportable as CSV. Scheduled email version available.

Commission Report
Agent commission ledger

Per-agent commission totals for selected period. Export as CSV for payroll. Shows tier thresholds and FTD counts.

Chargeback Log
All reversed deposits

Full chargeback history with client, amount, date, broker, and reason. Filter by broker to identify high-risk partners.

Export
Export current view

Exports the active filtered deposit log or P&L view as CSV or Excel. Logged in security audit trail with row count and filters applied.

P&L Dashboard
Full profit and loss view

Multi-dimension P&L with campaign, broker, shift, and agent breakdowns. Period comparison built in.

Anomaly Flags
Unusual activity queue

All auto-flagged deposits awaiting review. Each flag shows the rule triggered, deposit details, and a dismiss or escalate action.

Budget vs. Actual
Campaign spend tracker

Compares configured campaign budget against actual cost (ad spend + CPL). Red indicator when overspend detected.

Module 03

Chat

Real-time internal messaging powered by Socket.IO. Agents communicate in context — inside the CRM, not in external apps. Managers broadcast to the entire floor in one click.

Real-time via Socket.IO (Viki engine)
Read receipts and online presence
Floor broadcast to all agents simultaneously

Chat Types

TypeAccessible ToUse Case
Direct Message (DM)All usersOne-to-one communication between any two system users
Campaign ChannelUsers on that campaignTeam discussion space for a specific campaign — shared history, searchable
Manager BroadcastManager / Admin only (send)Single message delivered to ALL active agents — with read receipts
System NotificationSystem-generatedAuto-messages: task escalations, score drops, shift changes
Context sharing in DMsAgents can share a client profile directly in a DM — the recipient sees a clickable client card with name, status, and AI score. No copy-pasting needed.

Floor Broadcast — How to Send

1
Open Chat module

Navigate to Chat from the left sidebar. Manager/Admin role required to see broadcast option.

2
Click "Broadcast"

The broadcast compose window opens. Message will go to all agents currently active in the system.

3
Write message and send

Supports plain text. Keep broadcasts short and actionable. Recipients see it as a notification + in-chat message.

4
Monitor read receipts

The broadcast panel shows who has read the message. Unread count decreases as agents open the notification.

Chat is fully loggedAll messages — DMs, channel messages, and broadcasts — are stored in the system and visible in the Security audit log to admins. Do not use Chat for sensitive data outside of the system context.
Module 04

Security

Enterprise security at every layer. AES-256 session encryption, immutable audit logs, role-based access matrices, IP restrictions, and real-time anomaly detection.

AES-256 session encryption
Immutable audit log — write-once
Admin only access

The Audit Log

Every action performed by every user in the system is logged in the immutable audit trail. This log cannot be modified by any user, including admins. It is write-once, read-many. Each entry contains: timestamp, user ID, action type, affected record, IP address, and user agent string.

Event TypeTriggered ByWhat's Logged
Login / LogoutAny userUsername, IP, device, timestamp, success/failure
Client viewAgent, ManagerWhich client record was viewed, by whom, when
Client editAgent, ManagerBefore/after values of changed fields
ExportManager, AdminWhat was exported, filter parameters, row count
Permission changeAdminWhich user's role or permissions were changed, by whom
Failed loginUnauthenticatedIP address, attempted username, timestamp
Anomaly flagSystemTriggered event type, affected entity, severity level

Session Security

Sessions use AES-256 encryption with dual key scheme (APP_ENC_KEY_A + APP_ENC_KEY_B). The enc_user field stored in the database is an encrypted combination of the key pair and username — even direct database access does not expose plaintext user identities. Sessions auto-expire after the configured inactivity timeout.

Login uses username_special, not emailThis is intentional. Email addresses are known to attackers. The username_special is an internal identifier that provides one additional layer of obscurity against credential stuffing attacks.

Anomaly Detection Rules

Rapid failed logins
Flag
3+ failed login attempts from same IP within 5 minutes triggers IP block and admin notification.
Unusual export volume
Alert
Export of 1000+ records triggers manager notification and audit entry regardless of who performed it.
Off-hours access
Log
Access outside configured business hours is logged with elevated visibility in the audit report.
New IP login
Notify
First login from a new IP for any admin account triggers email notification to that admin.

IP Restriction Management

Admin can configure an IP whitelist that restricts system access to approved network ranges — for example, the office network's static IP block. Any login attempt from a non-whitelisted IP is blocked at the authentication layer before any credentials are validated.

IP whitelist
CIDR ranges
Comma-separated IP ranges in CIDR notation (e.g., 192.168.1.0/24). Configured in Settings → Security → IP Whitelist.
Role-specific restriction
Per-role
Agent logins restricted to whitelisted IPs only. Admins can optionally allow mobile access from any IP with an additional auth step.
Exception requests
Logged
Blocked login attempts from non-whitelisted IPs are logged with the attempted IP, username, and timestamp. Admin can approve exceptions individually.
IP restrictions apply immediately on saveWhen you add or remove an IP range from the whitelist, the change takes effect for all subsequent login attempts with no restart required. Existing active sessions are not terminated by whitelist changes — only new logins are affected.

Session Security — Technical Detail

Cowboy CRM sessions use dual-key AES-256 encryption. The session token stored in the database is an encrypted combination of APP_ENC_KEY_A and APP_ENC_KEY_B with the username — meaning even direct database read access by a compromised party does not expose who is logged in or what their session credentials are.

Session token
AES-256
Stored encrypted in the sessions table. Cannot be decoded without both key values.
Inactivity timeout
Configurable
Session expires after: 30 min / 1 hour / 4 hours / shift-end. Set in Settings → Security → Session Timeout.
Remote termination
Admin action
Admin can forcibly terminate any active session from the Users module. Session is invalidated immediately — the user is logged out within 30 seconds.
enc_user
Encrypted string
The stored session field. Contains encrypted identity — not a plaintext username. Used in routing and audit attribution.

Data Export Controls

Every export from Cowboy CRM is governed by role, volume, and rate limits. The goal is to prevent bulk data exfiltration while allowing normal operational reporting.

ControlRuleWhat Happens on Trigger
Role requirementManager role or above required to export any client dataExport button hidden for Agent role at UI and API level
Large export notificationExport of 500+ recordsAdmin receives notification with: who exported, when, filters applied, row count, IP address
Export log entryEvery export, regardless of sizeLogged in audit trail with full details. Cannot be deleted.
Rate limitingMax 3 large exports (500+ rows) per user per hourFurther exports blocked until the hour window resets. Admin can override for a specific user.

Viki — Real-Time Intelligence Engine (Technical)

Viki is the real-time statistics and event broadcasting system that powers live leaderboards, floor activity feeds, and instant routing notifications. It runs on MongoDB + Socket.IO alongside the main Laravel/MySQL application. Six live stat endpoints feed data to the frontend continuously:

Viki EndpointWhat It ProvidesConsumer
/viki/stats/floorLive floor overview: active agents, current call count, open leads, shift statusManager floor view, Viki widget
/viki/stats/agentIndividual agent live stats: current score, position on leaderboard, tasks dueAgent dashboard, gamification screen
/viki/stats/depositsLive deposit feed: new deposits, FTD count for today, deposit-in-progress alertsFinance module, manager view
/viki/stats/routingLive routing decisions: which leads were just routed, to which agents, and the AI score that drove the decisionAI routing log, admin view
/viki/stats/brokersLive broker performance: FTD count per broker in real time, pending approvals, chargeback alertsBrokers module live panel
/viki/stats/systemSystem health: API response time, queue depth, error rates, active Socket.IO connectionsAdmin → System Health panel
MongoDB + Socket.IO architectureViki uses MongoDB for fast real-time stat writes (high write throughput without locking the main MySQL database) and Socket.IO for push delivery to all connected browser sessions. No polling — all updates are server-pushed the moment data changes.

Automated Background Jobs (Cron)

Cowboy CRM runs a set of automated background jobs that maintain data integrity, calculate statistics, and run operations without any manual action required. These are the real cron jobs running in the production system:

Job NameScheduleWhat It Does
Retention JobsDaily at 2:00 AMProcesses client retention logic — identifies dormant high-value clients and creates re-engagement tasks for the retention team.
Statistics CompilationDaily at 2:00 AMCompiles daily, weekly, and monthly performance statistics for all agents, campaigns, and brokers. Feeds the Reports module with pre-aggregated data.
Affiliation ProcessingDaily at 2:00 AMCalculates affiliate CPA, fires postbacks and pixels for converted leads, updates affiliate balance ledger.
User OperationsDaily at 2:00 AMRecalculates composite scores for all active users, updates tier assignments, triggers scheduled gamification period rollovers.
Client ProcessingEvery minuteChecks for new leads arriving via broker API, scores and routes them immediately. Also processes scheduled callback tasks due in the next 5 minutes.
Task ProcessingEvery minuteChecks for overdue tasks and fires the escalation chain — Level 1 at 1 hour, Level 2 at 4 hours, Level 3 at 24 hours. Compliance tasks escalate immediately.
Affiliation Minute-CheckEvery minuteReal-time check for affiliate postback events that need immediate processing (fraud detection, cap limits, conversion fires).

User Deactivation Protocol

When an agent leaves the organization or must be removed from the system, use the Deactivate action — never delete. Deletion is not available for users who have any historical activity, ensuring full audit trail integrity.

1
Admin opens Users module

Navigate to the agent's profile. Click the "Deactivate User" action in the profile header.

2
Session terminated immediately

The agent's active session is invalidated within 30 seconds. They are logged out and cannot log back in.

3
Active leads redistributed

All leads currently assigned to the deactivated agent are auto-redistributed. Redistribution target is configurable: a specific named agent, or the AI auto-router.

4
Historical data preserved

All notes, tasks, call logs, and deposits attributed to this agent are preserved intact. The agent profile becomes read-only and visible to managers in the Inactive Users list.

You cannot delete users — only deactivateThis is by design. Deleting a user would remove attribution records, commission history, and audit trail entries. Deactivation preserves all historical integrity while preventing further access.
Module 13

Communications

AI-orchestrated outbound communications across email, SMS, and WhatsApp. The system selects channel, applies your template, and sends at the statistically optimal moment for each lead.

3 channels: Email, SMS, WhatsApp
AI-optimized send timing per lead
A/B testing built in

How AI-Optimized Timing Works

Cowboy CRM tracks the open rate, response rate, and conversion rate of every message sent — grouped by lead geography, campaign, time of day, and channel. Over time, it builds a statistical model of when each type of lead is most responsive on each channel. New sends are scheduled based on this model.

Example: German leads from finance campaigns open emails 3x more on Tuesday mornings than any other time. Cowboy CRM schedules those sends automatically without manager configuration.

Template Management

1
Navigate to Communications → Templates

The template library shows all saved templates organized by channel (Email / SMS / WhatsApp) and language.

2
Create or edit a template

Templates support dynamic fields: {{client_name}}, {{agent_name}}, {{callback_time}}, {{offer_title}}. These are populated automatically on send.

3
Set A/B variants (optional)

Create a B variant with a different subject line or body. System automatically splits traffic and reports winner.

4
Submit for compliance review

Templates go to draft state until approved by an admin or designated compliance reviewer. Approved templates go live immediately.

Communication Log — 11 Channel Types

Every interaction with a client is logged as a communication event with one of 11 real channel types. These appear in the client's timeline and in the Communications report. Agents select the type when logging a manual interaction; the system sets it automatically for system-generated events.

Channel TypeDescriptionCreated By
phonePhone call (inbound or outbound). Duration, outcome, and agent logged automatically.Agent (manual log) or system
email_sentOutbound email dispatched to the client. Linked to the template used and delivery status.System (from sequence) or agent
email_receivedInbound email reply received from the client. Auto-creates a task for the assigned agent.System (auto-detected)
skypeSkype call or message with the client. Agent logs manually after the interaction.Agent (manual)
chatInternal CRM chat interaction involving this client. Linked to the Chat module thread.System (from Chat module)
smsOutbound SMS sent to the client. Delivery status tracked. Reply creates email_received event.System (from sequence) or agent
whatsappWhatsApp message (sent or received). Read receipts tracked when available.Agent (manual) or system
viberViber message or call. Agent logs manually. Used in specific geographic markets.Agent (manual)
telegramTelegram message or call. Agent logs manually. Common in Russian-speaking and Asian markets.Agent (manual)
paypalPayPal interaction — payment confirmation, refund communication, or account verification message.Agent (manual)
btcBitcoin / crypto payment communication. Used when client deposits via crypto and confirmation is needed.Agent (manual)

Built-In Preset Messages

The system includes pre-written quick-send presets for both email and SMS — available for agents to dispatch in one click without opening the full template editor. These presets are designed for the most common agent-to-client situations.

6 Email Presets (Quick Send)

#Preset NameUse Case
1Waiting backAgent is following up on a previous call — "Still waiting to hear from you" message to re-engage a silent client.
2Just calledSent immediately after a call attempt to document the contact. Informs client agent tried to reach them.
3Great promotionPromotion-focused outreach. Highlights a current offer or bonus to incentivize the client to act.
4Come backRe-engagement message for dormant clients. Warm reminder to return to the conversation.
5Bad numberSent when the phone number on file is not working. Asks client to provide a corrected contact number.
6Still interestedSoft inquiry sent to gauge whether the client is still interested before investing agent time.

7 SMS Presets (Quick Send)

#Preset NameUse Case
1Thank youPost-call thank-you message. Sent after a positive call to reinforce the relationship.
2Just calledImmediate post-call SMS confirming agent attempted contact. Common follow-up to a missed call.
3Me againCasual re-contact. Agent letting client know they're still available and following up.
4Come backRe-engagement SMS for clients who have gone quiet. Short and direct.
5SupportSent when routing client to support. Informs them a support specialist will be in touch.
6Not InterestedAcknowledgment SMS when client has requested no further contact. Polite closure message.
7Not reachedCannot reach you — sent after multiple failed contact attempts. Asks client to call back or reply.

Do-Not-Contact Compliance

Clients with "Do Not Contact" status are automatically excluded from all outbound sequences — email, SMS, and WhatsApp. This exclusion is enforced at the API dispatch layer, not just the UI. A DNC-flagged client cannot receive system-generated communications even if manually included in a batch.

ChannelTracking AvailableA/B TestingDNC Enforced
EmailOpens, clicks, replies, bouncesYes — subject line and bodyYes
SMSDelivery confirmation, repliesYes — body onlyYes
WhatsAppDelivery, read receipts, repliesLimited — body variantsYes

Template Variables — Full Reference

Templates support dynamic variables that are replaced with real data at send time. Variables are case-sensitive and must use double curly brace syntax exactly as shown.

{{client_name}}
String
Client's first name as stored in the profile. Defaults to "there" if first name is blank.
{{agent_name}}
String
First name of the assigned agent. Personalizes the sign-off line of outbound messages.
{{callback_time}}
Formatted datetime
The scheduled callback datetime for this client, formatted in the client's local timezone (e.g., "Tuesday 14:30 CET").
{{offer_title}}
String
The campaign's configured offer name. Set in Settings → Campaigns → Offer Title.
{{broker_name}}
String
The broker's display name as configured in the Brokers module. Used in co-branded communications.
{{language}}
String
Client's configured language setting (e.g., "German", "French"). Useful for multilingual template logic.
{{country}}
String
Client's country as stored in the profile. Used in localization and geo-specific messaging.
{{ai_score}}
Integer
Client's current AI score (0–100). For internal emails only — never use in client-facing templates. Intended for manager digest emails.

A/B Testing Engine

Every template can have a B variant. The A/B test engine automatically splits sends between variants and tracks results. After a configurable sample threshold is reached, the system declares a winner and continues sending only the winning variant — no manual intervention required.

1
Create variant B

In the template editor, click "Add B Variant." Write the alternative subject line and/or body. Variants can differ in subject only, body only, or both.

2
Set split ratio and sample size

Default: 50/50 split. Configurable to 70/30 or any ratio. Set the minimum sample size (e.g., 200 sends per variant) before a winner is declared.

3
Set winner criteria

Choose what defines the winner: open rate (email), click rate (email), or reply rate (email, SMS, WhatsApp). Configurable per test.

4
System auto-selects winner

Once sample size is reached, the winning variant is automatically promoted to 100% of sends. Test history and result data are preserved in the template record.

A/B test history is cumulativePrevious test results for each template are stored in the test history panel. Use them to build institutional knowledge about what messaging works for each language and campaign segment.

Send Timing Optimization

Cowboy CRM builds a send timing model per lead segment from historical data. The model groups leads by country + campaign + day-of-week and tracks open rates and reply rates for each hour of the day within those groups.

Segment ExamplePeak Send Window (Example)Basis
German finance leads — Tuesday09:00 – 11:00 CETHistorical open rate analysis across 6-month window
Spanish retail leads — Weekend19:00 – 21:00 CETReply rate peak derived from 3,000+ sent messages
New segment (insufficient data)Falls back to campaign default windowUses campaign-level average until 100+ sends in that segment

Scheduled sends use the peak window for the recipient's segment automatically. Manager override is always available: click "Send Now" on any scheduled message to bypass timing optimization and dispatch immediately.

DNC (Do-Not-Contact) Management

The DNC system enforces compliance at the API dispatch layer — not just the UI. A client flagged as DNC cannot receive system-generated communications even if a manager manually adds them to a batch. The block is architectural, not a UI checkbox.

Do Not Call
Status flag
Excludes client from outbound call task creation. Agents will see a DNC warning on the client profile.
Do Not Email
Status flag
Blocks email dispatch at the API layer. Client is skipped silently in bulk sends with a logged skip reason.
Do Not SMS
Status flag
Blocks SMS dispatch. Applied separately from email — a client can be email-OK but SMS-blocked.
Do Not Contact (all)
Status flag
Global block across all channels. Sets all three channel flags simultaneously. Shown as a red banner on the client profile.

DNC can be set in three ways: manually by a manager on the client profile, via bulk import (upload a CSV of phone numbers or emails to flag in bulk), or automatically — when a delivery failure plus opt-out reply combination is detected for email, the system sets email DNC automatically without agent action.

DNC log is auditableEvery DNC flag — who set it, when, and why — is stored in the DNC log accessible in Security → Audit Log. Required for regulatory compliance review.

Delivery Tracking — Per Channel

ChannelEvents Tracked
EmailSent → Delivered → Opened → Clicked → Replied → Bounced → Unsubscribed
SMSSent → Delivered → Reply received → Failed
WhatsAppSent → Delivered → Read → Replied → Failed

All delivery events are tracked per client record and in aggregate reports. The client timeline shows the delivery status of every message sent. The Communications report aggregates open rates, click rates, bounce rates, and reply rates across all sends for a given template, campaign, or date range.

Module 08

Brokers

Unified broker intelligence. 37+ brokers tracked in real time. FTD attribution, conversion scoring, and automatic lead rebalancing — all from one interface.

37+ brokers managed
Auto-rebalancing on underperformance
API integration with major broker platforms

Broker Performance Scoring

Every broker in the system receives a real-time performance score based on the quality of leads they send. This score is used by the routing engine to prioritize which broker's leads receive premium agent assignment. Underperforming brokers are flagged for review automatically.

MetricWeightHow Measured
Lead-to-FTD conversion rate35%FTDs attributable to this broker / total leads from broker
Revenue per lead30%Total attributed deposits / total leads delivered
Lead freshness20%Average time between lead creation and first contact attempt
Chargeback rate15%Disputed or reversed deposits / total deposits attributed

Automatic Lead Rebalancing

When a broker's performance score drops below the configured threshold (set in Settings), the system automatically reduces lead allocation from that broker. The threshold, reduction percentage, and notification recipients are all configurable. No manual intervention required — the system adjusts continuously.

Rebalancing is not a banAutomatic rebalancing reduces allocation — it does not stop lead flow entirely. Admins can set a floor allocation percentage per broker to prevent full cutoff. A complete ban requires a manual status change to Inactive.

Broker Fields

broker_name
Text
Display name. Used in all reports and attribution labels.
api_endpoint
URL
If using API integration, leads push here automatically when status updates occur in Cowboy CRM.
performance_score
Float 0-100
Auto-calculated. Recalculated every 24 hours based on trailing 30-day data.
lead_allocation_pct
Percentage
What fraction of total incoming leads are attributed to this broker. Adjusted by auto-rebalancing.
status
Enum
Active / Paused / Inactive. Inactive brokers' leads still exist in the system but receive no new routing priority.
campaigns
Multi-select
Which campaigns this broker's leads are eligible for. Determines routing pool membership.

Broker API Integration

Cowboy CRM maintains a two-way API connection with integrated brokers. On the inbound side, brokers push new leads directly to Cowboy CRM via POST. On the outbound side, Cowboy CRM pushes status updates back to the broker's webhook URL whenever key events occur.

Outbound Status Push Events

First contact attempt
Push
Fired when an agent makes the first call attempt on a broker's lead. Signals to broker that the lead has been worked.
Deposit recorded
Push
Fired on FTD confirmation. Triggers broker-side commission tracking and affiliate postback chain.
Rejection
Push
Fired when client status is set to Rejected or Do Not Call. Allows broker to update their own CRM.
Callback scheduled
Push
Optionally fired when a callback is scheduled — available for brokers whose systems track contact attempt timing.

API authentication uses token-based auth per broker endpoint. Each broker's API token is configured in the broker profile under Settings. API health — response codes, last successful push, error count — is visible in Settings → System Health panel.

Failed API pushes are retried automaticallyIf a broker's webhook returns an error, Cowboy CRM queues the push for retry with exponential backoff (3 retries over 30 minutes). Persistent failures are flagged in the API health panel with the error code and last attempt timestamp.

Lead Allocation Management

Total incoming lead volume is distributed across active brokers by configurable percentage. The sum of all active broker allocations must equal 100%. Auto-rebalancing adjusts these percentages automatically based on performance scores — but admin can lock any broker's allocation to prevent automatic changes.

SettingWhat It ControlsWhere to Configure
Allocation %Share of total incoming leads routed to this brokerBroker profile → Lead Allocation field
Floor allocationMinimum % auto-rebalancing will not go below — prevents full cutoffBroker profile → Floor Allocation field
Auto-rebalance lockWhen enabled, admin controls this broker's allocation manually; auto-rebalancing is disabled for this broker onlyBroker profile → Lock Allocation toggle
Rebalance thresholdPerformance score below which auto-reduction triggersSettings → Brokers → Rebalance Threshold
Rebalancing is not a banAutomatic rebalancing reduces allocation — it does not stop lead flow entirely. A complete ban requires a manual status change to Inactive in the broker profile.

Broker Comparison Dashboard

The comparison dashboard presents all active brokers side-by-side in a single table, sortable by any column. It is the primary screen for contract renegotiation preparation and quarterly broker performance reviews.

ColumnDefinition
Leads SentTotal leads delivered from this broker in the selected date range
FTDsFirst-time deposits attributable to this broker's leads
Conversion RateFTDs / Leads — the primary lead quality indicator
Revenue GeneratedTotal deposit value from this broker's leads (base currency)
CPA PaidTotal agreed cost paid to broker for the FTDs in this period
Net ValueRevenue Generated minus CPA Paid — the actual profit contribution of this broker
Performance ScoreCowboy CRM's composite score (0–100) based on conversion rate, revenue/lead, freshness, and chargeback rate

Click any column header to sort ascending or descending. Apply a date range filter (last 7 / 30 / 90 days) to scope the comparison. Export the current view as CSV for contract renegotiation meetings — the export includes all columns and the applied date range as a header row.

Reference · Affiliate Tracking

Affiliation

Click-to-close attribution for every traffic source. Affiliate CPA calculation, channel ROI comparison, traffic quality scoring, pixel tracking, and fraud signal detection.

Click-level source attribution
CPA auto-calculated per affiliate
Fraud signals flagged automatically

Attribution Model

Every lead that enters Cowboy CRM carries a source tag set at the moment of creation: affiliate ID, campaign, landing page, keyword, and traffic channel. This tag is immutable — it cannot be changed after the lead is created. When the lead converts and deposits, the attribution flows back automatically to the originating affiliate.

Postback / pixel firingOn FTD conversion, Cowboy CRM fires the postback URL or pixel configured for the originating affiliate automatically. No manual reporting to affiliates needed.

Affiliate Performance Dashboard

ColumnDefinition
LeadsTotal leads sourced from this affiliate in the selected period
FTDsFirst-time deposits attributable to this affiliate's leads
Conv. RateFTDs / Leads — the primary quality indicator
RevenueTotal deposits attributed, in base currency
CPATotal affiliate cost / FTDs — how much each conversion cost
Quality ScoreComposite of conversion rate, lead freshness, and chargeback rate
Fraud SignalsCount of leads flagged for duplicate phone, bot pattern, or incentivized behavior

Fraud Detection Signals

Duplicate submission
Flag
Same phone or email submitted multiple times from same affiliate within 24 hours.
Completion-time anomaly
Flag
Form completed in under 3 seconds — indicates automated or pre-filled submission.
Geographic mismatch
Alert
IP country does not match declared country — common in incentivized traffic operations.
Zero-conversion cluster
Alert
50+ leads from same affiliate with 0 FTDs triggers automatic quality review flag for admin.
Reference · Training Feature

Quiz

Continuous knowledge testing for agents. Compliance quizzes, product tests, and sales training — all scored, tracked, mandatory where configured, and factored into agent performance profiles.

Automated scheduling by role and campaign
15% weight in agent composite score
Escalation on failed compliance tests

Quiz Types

TypeCreated ByMandatoryEscalation on Fail
Compliance QuizAdmin / Compliance officerYesYes — team lead notified, agent status flagged
Product KnowledgeManager / AdminPer campaignWarning on first fail, escalation on repeat
Sales TrainingManagerNoNo — logged for coaching reference only
Onboarding QuizAdminYes — new usersYes — new user cannot access leads until passed

Creating a Quiz (Manager / Admin)

1
Navigate to Quiz → Create

The quiz builder opens. Set title, description, passing score threshold (% correct), and whether it is mandatory.

2
Add questions

Multiple choice, true/false, or scored open-ended formats. Set correct answers and point values per question.

3
Assign audience

Assign to all agents, agents on specific campaigns, specific roles, or individual users. Time limit is optional.

4
Set recurrence (optional)

Compliance quizzes can repeat monthly or quarterly automatically. System creates new quiz assignment on schedule.

5
Publish

Assigned agents see the quiz in their task list immediately. Mandatory quizzes show a banner until completed.

Failed mandatory compliance quizIf an agent fails a mandatory compliance quiz, their account is flagged. A manager notification is sent. The agent may continue working but the flag is visible to their team lead and appears in their performance profile.

Quiz Scores in Routing

Quiz scores contribute 15% of the agent composite score. An agent who consistently scores high on compliance and product quizzes is rated more reliable for high-value leads. An agent with recent failures on campaign-specific quizzes is automatically downweighted in routing for that campaign until scores recover.

Reference · System Configuration

Settings

The complete system configuration interface. Role builder, module visibility, routing parameter tuning, API integrations, notification preferences, and system health monitoring — all admin-only.

Admin only — restricted at API level
All changes logged in audit trail
Live health panel for integrations
Admin role requiredSettings is completely hidden from Agent, Manager, and Viewer roles at the API level. Accessing Settings API endpoints without admin authentication returns 403.

Settings Sections

SectionWhat You Configure
Roles & PermissionsCreate custom roles. Grant/deny module access per role. Set field-level restrictions.
Routing ParametersAdjust AI routing weight factors: language match weight, load balancing factor, performance history window.
IntegrationsConfigure broker API endpoints, affiliate postback URLs, SMS gateway, email provider, WhatsApp API key.
Module VisibilityShow/hide entire modules from specific roles. Finance can be hidden from managers. Games can be disabled organization-wide.
NotificationsConfigure which events trigger which notifications (email, SMS, in-system) to which role or specific user.
FX RatesSet currency conversion rates for multi-currency Finance reporting. Updated manually or via auto-sync.
Security PolicySession timeout duration, IP whitelist rules, failed login lockout threshold, anomaly alert recipients.
System HealthLive status of all integrations: database, Socket.IO engine, SMS gateway, email provider, broker APIs.

Routing Parameter Tuning

The routing algorithm uses configurable weights for each scoring factor. The defaults are calibrated from years of call center data, but every operation is different. Admins can increase the weight of language match if language mismatches are causing churn, or increase the performance history weight if high-scoring agents are being underutilized.

Change one parameter at a timeRouting changes affect the entire lead distribution immediately. Make incremental adjustments and monitor Statistics for 24–48 hours before making further changes. Document what you changed and why in the admin notes field.

System Health Panel

The health panel shows real-time status for: database connection and query response time, Viki (Socket.IO) engine status and connected clients, email provider delivery rate, SMS gateway delivery rate, and each configured broker API's last successful response. Any degraded component shows in amber; a failed component shows in red with the last successful contact timestamp.

Getting Started

Login & Authentication

Cowboy CRM uses encrypted username-based authentication. Credentials are provided by your admin. Sessions expire after inactivity.

Login Process

1
Navigate to cowyboymanager.com

The login screen appears automatically. Do not bookmark internal pages directly.

2
Enter your username_special

This is NOT your email address. It is a unique username provided by your admin on account setup.

3
Enter your password

Passwords are AES-256 encrypted. If you forget your password, contact your admin — there is no self-reset by design.

4
System loads your view

Based on your role, you see your personalized dashboard. Agents see their leads. Managers see the floor overview.

Getting Started

Navigation

The Cowboy CRM interface uses a left sidebar for module navigation and a top bar for filters, charts, calendar, Viki (AI), and user settings.

Top Bar Elements

Filters
Global filter panel

Apply filters that affect the current module view — date, agent, broker, campaign, status.

Charts
Chart overlay

Toggles chart view on top of the current data. Available in Clients, Statistics, Shifts.

Calendar
Calendar view

Switch between list and calendar layout for tasks, shifts, and scheduled calls.

Viki AI
Viki — the AI assistant

The real-time AI engine visible from every screen. Shows live routing decisions, queue status, and system alerts.

Reference

The AI Engine

The PCT-patented intelligence layer shared by Cowboy CRM and SortExpress.com. The engine that makes every routing decision, timing call, and performance score.

How the Algorithm Works

The scoring algorithm was designed in 2017 by Guy Evgeni Levin and is protected under PCT international patent. It processes 40+ parameters per lead simultaneously to generate a composite score that drives all routing decisions.

Parameter CategoryExamplesWeight Impact
Geographic signalsCountry, city, timezone offset, regionHigh
Language matchBrowser language, form language, detected localeHigh
Source qualityTraffic source, affiliate score, broker ratingHigh
Temporal signalsTime of day, day of week, holiday calendarMedium
Device & behaviorDevice type, form completion time, field interactionsMedium
Historical dataSimilar lead outcomes from same source/geo/timeHigh
Agent availabilityCurrent shift, load, language, recent performanceHigh
Reference

API & Integrations

Cowboy CRM exposes a REST API for all modules. The SortExpress integration is native. Custom broker integrations via webhook or direct API.

SortExpress integration is built-inNo configuration needed. If your account uses SortExpress, leads flow directly into Cowboy CRM via secured API with full attribution intact.
Reference

Frequently Asked Questions

Common questions from agents, managers, and admins across all 16 enterprise clients.

For Agents

Why was this lead assigned to me and not another agent?

Cowboy CRM scored this lead and matched it to you based on your language capabilities, current shift, load at the time of assignment, and your performance history with similar leads.

My score dropped — why?

Scores update in real time. A missed callback, a late clock-in, or a failed quiz can reduce your score. Check the score breakdown in your profile for the specific factor.

Can I see other agents' leads?

No. Agents only see leads assigned to them. This is by design and enforced at the database level, not just the UI.

For Managers

Can I override an AI routing decision?

Yes. In the client profile, click "Reassign" to manually move a client to a different agent. The system logs the override and uses it in future model calibration.

Does the system work if I do nothing?

Yes. That is the design. Shifts are built, leads are routed, calls are scheduled, emails are sent — all without manager input. Your job is to monitor performance and intervene when strategic decisions are needed.

Module 11

Notifications

Intelligent alert system that keeps every stakeholder informed in real time. System events, performance alerts, compliance triggers, and manager escalations — all routed automatically to the right person through the right channel.

Zero-config by default
3 channels: in-system, email, SMS
Role-targeted alerts

How Notifications Work

Cowboy CRM generates notifications automatically from system events — no manual configuration required for standard alerts. Each notification is routed to the correct recipients based on role: agents receive task and lead alerts; managers receive performance and escalation alerts; admins receive security and system health alerts.

Smart routing, not spamNotifications are suppressed during active shifts if the agent is in a call. They queue and deliver between calls, preventing interruptions at critical moments.

Notification Types

EventWho Gets ItChannelWhen
New lead assignedAssigned agentIn-systemImmediately
Task overdueAgent + team leadIn-system + email1 hour after deadline
Agent score dropManagerIn-systemDaily summary or threshold breach
Anomaly detectedAdminIn-system + emailImmediately
Quiz assignedAgentIn-systemOn assignment
FTD milestoneAll agents + managerIn-systemOn deposit
Shift starting in 30minAssigned agentsIn-system + SMS (if configured)30 min pre-shift
Broker performance dropAdminEmailDaily digest
System health alertAdminEmail + SMSImmediately on failure

Configuring Notifications (Admin)

Default notification rules work out-of-the-box. Admins can adjust thresholds, add additional recipients, change delivery channels, or disable specific event types in Settings → Notifications. Custom notification rules can be created for any system event with conditional logic.

Best practice: night-time suppressionConfigure the notification quiet window (e.g., 23:00–07:00) in Settings so system health alerts don't wake up managers for non-critical events. Critical security alerts bypass the quiet window regardless.
Module 12

Calendar

The operational calendar layer. Shift schedules, callback commitments, campaign windows, agent availability, and system events — all unified in one time-aware view that connects directly to the AI routing engine.

Shift schedule auto-populated
Callback queue visualized by time
Campaign windows color-coded

What's on the Calendar

Calendar ItemCreated ByLinked To
Shift blocksAI auto-generated or manager-manualShifts module — click to open shift details
Callback commitmentsAuto from client status changesClient profile — click to open client + task
Campaign windowsAuto from campaign settingsCampaign configuration
Quiz deadlinesAuto when quiz assigned to userQuiz module
Manual eventsManager or AdminStandalone — notes only
Calendar is read + writeClicking a callback item on the calendar opens the client profile directly. Clicking a shift block opens the shift editor. Everything is clickable — it is not a passive display.

View Modes

Day View
Hour-by-hour today

Shows all shifts, callbacks, and events for today. Agents see only their own items. Managers see the full floor.

Week View
7-day rolling window

Full week overview with shift density indicators. Useful for shift planning and spotting coverage gaps.

Callback Queue
Callbacks sorted by time

Lists all committed callbacks chronologically. Shows which are overdue (red), due within 30min (gold), and upcoming (blue).

Agent Timeline
Per-agent daily schedule

Shows one agent's full day: shift, leads assigned, callbacks due, tasks. Manager view only.

Module 14

Documents

Centralized document management linked directly to client records. KYC documents, signed agreements, call recordings, compliance certificates, and operational templates — all stored, versioned, and access-controlled.

Per-client document vault
Access-controlled by role
Audit-logged every view and download

Document Types

TypeLinked ToWho Can Access
KYC / ID documentsClient profileManager, Admin
Signed agreementsClient profileManager, Admin
Call recordingsClient profile + specific taskManager, Admin
Compliance certificatesUser profileAdmin
Operational templatesGlobal libraryManager, Admin
Training materialsGlobal libraryAll roles
Broker agreementsBroker profileAdmin

Document Workflow

1
Upload from client profile

Navigate to a client profile → Documents tab. Drag and drop or browse to upload. System records uploader, timestamp, and file hash.

2
Categorize and tag

Select document type, add optional description. Tags enable search across all documents in the system.

3
Set access level

Default access follows role permissions. Documents can be locked to Admin-only or shared with the client's assigned agent.

4
Access logged automatically

Every view and download is logged in the security audit trail: who, when, IP address, file name.

Documents cannot be deleted — only archivedFor compliance and audit integrity, documents uploaded to client records are immutable. They can be archived (hidden from normal view) but the record of their existence is permanent.
CEO & Buyers

CEO FAQ

Every question a CEO, VP of Sales, Operations Director, or IT Manager asks before making a purchasing decision. Real answers about Cowboy CRM's real capabilities, architecture, and outcomes.

50+ questions answered
7 categories: fit, AI, ROI, implementation, security, comparisons, operations
Specific facts, not marketing language

Is Cowboy CRM Right For My Operation?

What size call center is Cowboy CRM built for?

Cowboy CRM was designed and tested in operations ranging from 8 agents on a single campaign to 120+ agents across multiple simultaneous campaigns in 6 languages. The sweet spot is 15–150 agents where the cost of human routing decisions is measurable and the complexity exceeds what a manager can handle manually. Below 10 agents, much of the automation is overkill. Above 200, contact the team about enterprise deployment architecture.

What industries use Cowboy CRM?

Cowboy CRM was originally built for the financial services call center vertical — specifically operations handling investment and trading leads at high volume. The 16 current enterprise clients are in this sector. The core system — AI routing, shift management, gamification, broker attribution — applies to any outbound or inbound call center where lead quality varies by source, agent-to-lead matching matters, and conversion data drives operations.

We already have a CRM. Why replace it?

Generic CRMs (Salesforce, HubSpot, Monday.com) are record systems. Cowboy CRM is a decision engine. The difference: a CRM stores data and waits for humans to act on it. Cowboy CRM analyzes data and acts autonomously — routing leads in 300ms, building next week's shifts tonight, sending the right email at the right moment without a manager configuring a workflow. If your current CRM requires a manager to decide who gets each lead, you are paying for human decisions that an algorithm can make better.

We operate in multiple countries and languages. Does Cowboy CRM handle this?

Multi-country, multi-language operations are exactly what Cowboy CRM was built for. The routing engine considers language as a primary parameter — a German-speaking lead will never route to an agent who only speaks English. Time zone, local holidays, country-specific calling hours, and currency are all handled per campaign. The system currently serves clients operating across European, Middle Eastern, and LATAM markets simultaneously from single deployments.

We have 168,000 leads in our existing system. Can we migrate?

Yes. The Cowboy CRM database currently holds 168,000+ client records from production use. CSV import is the standard migration path. The import process maps your existing fields to Cowboy CRM's schema, preserves history notes, and runs a duplicate-phone check before committing. The AI engine begins scoring imported leads immediately on import completion — no warm-up period required.

Does Cowboy CRM work for inbound as well as outbound operations?

The core system was built for outbound operations where lead assignment is the primary routing challenge. For inbound operations, the Calendar, Shift, and Notification modules provide full coverage. The AI routing engine can also route inbound calls to the best-matched available agent based on the calling number's country, language, and historical interaction profile if a CRM record exists.

What if our operation has unique compliance requirements?

The Security module's role-based access controls, Do-Not-Contact enforcement at the API layer, and immutable audit logs address most compliance requirements in regulated industries. The Settings module allows admins to configure which fields are visible to which roles, enforce data retention policies, and set IP whitelists. For specific regulatory frameworks, speak with the implementation team.

How many brokers can the system manage simultaneously?

The Brokers module currently manages 37+ active brokers in production for a single enterprise client. There is no hard limit. Each broker has independent performance scoring, allocation percentages, API endpoint configuration, and campaign assignments. The auto-rebalancing system handles load distribution across all active brokers without manual intervention.

How Does the AI Actually Work?

What makes Cowboy CRM's AI different from other "AI-powered" CRMs?

Most CRM vendors apply "AI" to mean a recommendation engine that suggests next actions, or a chatbot. Cowboy CRM's algorithm is an autonomous routing system that processes 40+ parameters in under 300ms and makes binding operational decisions — not suggestions. It was designed from the ground up as the operating brain of a call center, not as an add-on feature to a contact management system. This architecture is protected under PCT international patent, filed by Cowboy CRM's creator.

What are the 40+ parameters the algorithm processes?

The parameters fall into 7 categories: geographic signals (country, city, timezone, region), language match (browser language, form language, declared language), source quality (broker rating, affiliate score, traffic channel quality), temporal signals (time of day, day of week, local holidays), device and behavioral signals (device type, form completion time, field interactions), historical patterns (similar lead outcomes from same source/geo/time slot), and agent state (current shift, load, language capabilities, composite performance score). All parameters are weighted, scored, and combined into a single routing decision per lead.

How does the system learn over time?

Every outcome — converted, rejected, no answer, callback — feeds back into the scoring model. When a lead from Source A in Country B assigned to Agent C at Time D converts at a high rate, that pattern is weighted more heavily for future similar leads. The model does not require manual retraining — it updates continuously from production data. After 6 months of operation, routing decisions are significantly more refined than at deployment.

Can managers override AI decisions? Does overriding hurt the model?

Yes, managers can override any routing decision by manually reassigning a client in the client profile. Overrides are logged with a timestamp and reason note. The override is treated as a data point — if managers consistently override a certain pattern, the system treats this as signal that the algorithm's weighting needs adjustment for that segment. Overrides do not damage the model; they improve it.

How fast is the routing decision? Does it slow down at scale?

Routing decisions are made in under 300ms regardless of the database size. The scoring calculation runs in application memory — it does not query the full client database for each decision. Performance has been validated at 168,000+ records without degradation. At very high volume (1000+ simultaneous lead arrivals per minute), the system queues leads and processes in order of arrival, maintaining sub-second decisions throughout.

What does "optimal agent" mean, exactly?

Optimal means the agent most likely to convert this specific lead right now, given current conditions. That is a combination of: language match (highest weight), current shift status (is the agent on shift?), current load (how many active leads does the agent already have?), recent performance with similar lead type, and composite performance score. An agent with a high score but wrong language will never receive a lead outside their language set. An on-shift agent always beats an off-shift agent regardless of score.

Can the AI decide to pause a broker automatically?

The system performs automatic lead rebalancing — reducing allocation from underperforming brokers — but does not fully pause or ban a broker without admin confirmation. When a broker drops below the configured performance threshold, allocation is reduced by a configurable percentage and an admin alert is sent. Complete deactivation requires a manual status change to prevent automated shutdowns affecting contractual broker relationships without human sign-off.

Is the AI algorithm the same one powering SortExpress?

Yes. The scoring and routing algorithm is the shared engine behind both Cowboy CRM and SortExpress.com. When an operation uses both platforms, leads generated by SortExpress arrive in Cowboy CRM pre-scored — the same algorithm that qualified the lead also routes it, creating a seamless qualification-to-routing pipeline with no handoff gap.

Can I see what the AI decided and why?

Yes. Every client profile includes a routing log showing every AI decision made for that client: the timestamp, the agents considered, the scores assigned, and the routing outcome. This log is visible to managers and admins. It is the primary tool for understanding and auditing routing behavior on specific leads.

What ROI Should I Expect?

What's the typical improvement in FTD conversion rate after deployment?

Based on production data across 16 enterprise clients, the primary gains come from three sources: eliminating language mismatch (leads no longer assigned to agents who cannot communicate with them), reducing time-to-first-contact (leads routed in 300ms vs. minutes in a manual system), and shift optimization (highest-converting hours are staffed by the best-matched agents). Operations that were manually routing leads typically see measurable conversion improvement within 30–60 days of deployment as the model accumulates data.

Where does the ROI come from operationally?

There are three categories. First, conversion rate improvement from better routing — every lead matched to the right agent at the right time. Second, manager time recaptured — managers in manual operations spend significant time on routing decisions that Cowboy CRM handles automatically, freeing them for coaching and strategy. Third, shift efficiency — Cowboy CRM's shift building is data-driven, meaning the right number of agents are on shift during high-conversion windows and overstaffing during low-conversion windows is reduced.

What's the payback period for a typical deployment?

This depends on operation size and current conversion rate baseline. In operations with 20+ agents, where each 0.1% improvement in FTD rate translates to meaningful revenue, payback periods of 2–4 months are typical. The gamification system (leaderboards, badges, performance scoring) has an additional documented effect on agent motivation and retention that reduces hiring and training costs over time.

How does the gamification system affect agent performance?

The leaderboard creates visible competition among agents updated in real time via Socket.IO. Agents on a floor that can see who is ranked first and what they need to do to overtake them exhibit measurably different behavior than agents on a floor with only a manager's verbal feedback. The badge system creates milestone events that broadcast across the floor — an FTD milestone announcement creates peer pressure and motivation simultaneously. Quiz scores feeding into routing also create an indirect incentive: agents who score better on product knowledge receive better-quality leads.

Does the system generate reports automatically that I can share with investors or board members?

Yes. The Reports module generates scheduled summaries — daily FTD counts, weekly agent performance rankings, monthly campaign P&L — automatically and delivers them via email. The Viewer role was specifically designed for C-level executives and investors who need read-only access to dashboards and summary reports without operational access. Reports can be configured to be emailed on schedule as PDF or CSV attachments.

Can we track which campaigns and brokers are profitable and which are losing money?

This is a core function of the Finance module. Every deposit is attributed across five dimensions simultaneously: agent, campaign, broker, traffic channel, and shift slot. Campaign P&L reports show revenue minus configured cost per campaign in real time. Broker comparison reports show revenue per lead, CPA, and chargeback rate side by side across all active brokers. Unprofitable campaigns and low-quality brokers are visible immediately.

Implementation & Migration

How long does deployment take?

A standard deployment — server setup, database configuration, user account creation, campaign setup, and broker API connections — takes 1–2 weeks for a technically capable team following the implementation guide. Data migration from an existing CRM adds 2–5 days depending on the volume and quality of the source data. The first AI routing decisions begin immediately after the first leads are imported and agents are on shift.

What technical resources do we need for deployment?

You need a server (the system runs on Node.js + MySQL; Cowboy CRM's standard deployment targets a VPS with 4+ cores and 8GB+ RAM), a domain, and an admin-level person to manage initial configuration. No specialized AI expertise is required — the algorithm is configured through the Settings UI, not through code. Integration with broker APIs requires knowledge of your brokers' API documentation. SortExpress integration is automatic with no setup required.

Can we run Cowboy CRM alongside our existing CRM during transition?

Yes. Because Cowboy CRM has API endpoints for all major entities, it can receive leads from existing systems during a parallel-run period. Many clients run Cowboy CRM as the routing and shift management layer while keeping legacy systems for historical record access during the first 30–60 days. Full replacement is recommended once agents are comfortable with the new system and historical data has been migrated.

How do we set up broker API integrations?

Broker integrations are configured in Settings → Integrations. For each broker, you configure: the inbound endpoint URL (where Cowboy CRM receives leads from the broker), the outbound API endpoint (where Cowboy CRM sends status updates back to the broker), authentication credentials, and campaign mapping. Cowboy CRM supports standard REST + JSON broker APIs. Custom webhook formats are handled during the implementation phase with support from the Cowboy CRM team.

How are agents trained on the new system?

Agent training is minimal because agents interact with a small subset of the system: their lead list, task queue, chat, and personal score/leaderboard. The YouTube walkthrough series (available at the EvolutionCRM channel) covers agent-specific workflows in under 30 minutes of video. For managers, the full system training typically takes 1–2 days. Admins handling configuration should spend 2–3 days with this guide and the implementation documentation before going live.

What happens to our data if we stop using Cowboy CRM?

All data is stored in your MySQL database. A full database export can be provided in standard SQL format at any time. Client records, interaction history, deposit logs, audit trails, and documents are all yours. There is no proprietary data format that prevents migration out — the database schema is documented and available to your technical team.

Security & Compliance

How is user data protected?

Sessions use AES-256 encryption with a dual-key scheme (APP_ENC_KEY_A + APP_ENC_KEY_B). The enc_user field stored in the database is an encrypted combination of the key pair and username — even direct database access does not expose plaintext user identities. All API calls require valid session tokens. The database is not exposed to the internet; all access goes through the application layer.

Is there a full audit trail of who accessed what?

Yes. The Security module maintains an immutable audit log. Every action by every user is recorded: login/logout events with IP and device, client record views, client edits with before/after field values, data exports with filter parameters and row counts, permission changes, and failed login attempts. This log is write-once and cannot be modified by any user including admins. It is accessible only to admin-role users.

How does Cowboy CRM handle GDPR "Do Not Contact" requirements?

Clients with "Do Not Contact" status are excluded from all outbound communications — email, SMS, WhatsApp, and call scheduling — at the API dispatch layer. This is not a UI checkbox that can be bypassed; it is enforced in the API code before any communication is dispatched. DNC flags are immutable once set by a client or compliance admin. The audit log records who set the DNC flag and when, providing the documentation trail required by regulators.

Can we restrict which agents can see which clients?

Yes. By design, agents only see clients assigned to them. This restriction is enforced at the database query level, not just the UI. If an agent's session is used to query the API for another agent's clients, the query returns zero results — not a permission error. Manager and admin roles see all clients by default, with optional campaign-based filtering available.

Can we limit access to specific IP addresses?

Yes. IP whitelist rules are configured in Settings → Security Policy. You can restrict all admin logins to office IPs or a VPN range. New IP logins for admin accounts trigger email notifications regardless of whitelist configuration. IP blocking is also applied automatically after 3+ failed login attempts from the same IP within 5 minutes.

Where is data stored? Can we self-host?

Cowboy CRM is designed for self-hosting on your own server infrastructure. The standard deployment is on a dedicated VPS (Hetzner, DigitalOcean, AWS, etc.) under your control. Your data does not pass through Cowboy CRM's servers — the software runs on your server against your database. This is a significant advantage for organizations with data sovereignty requirements.

Vs. Other CRM Systems

How is Cowboy CRM different from Salesforce?

Salesforce is a general-purpose CRM built for sales pipeline management across any industry. It is a powerful record system with a marketplace of integrations. Cowboy CRM is a purpose-built call center operating system. Salesforce does not build shifts, does not route leads autonomously based on 40+ parameters, does not have a gamification engine tied to real-time performance data, and does not have broker lead attribution with auto-rebalancing. You can build some of this in Salesforce with significant customization and cost — or use a system where it is the core function.

How is Cowboy CRM different from HubSpot?

HubSpot excels at inbound marketing automation and contact management for mid-market B2B sales teams. It is not designed for high-volume outbound call centers with shift-based operations. HubSpot has no concept of agent shifts, no real-time floor leaderboards, no broker performance scoring, and no autonomous lead routing engine. Its "sequences" feature automates email outreach — Cowboy CRM's Communications module does this while also routing calls, building shifts, and scoring agents simultaneously.

What about call center-specific platforms like Five9 or RingCentral?

Five9 and RingCentral are telephony platforms — they handle the actual calling infrastructure, IVR, queue management, and call recording. Cowboy CRM is the CRM and AI routing brain above the telephony layer. They solve different problems. Cowboy CRM integrates with telephony platforms via API — the call is made through your telephony system, and Cowboy CRM handles lead management, routing, shift scheduling, compliance, and analytics.

What about other "AI CRM" systems on the market?

Most systems marketed as "AI CRM" use AI for one specific function: a chatbot, a next-best-action recommendation, or a predictive lead score displayed as a number. Cowboy CRM's AI is the routing engine itself — not a layer on top of a CRM, but the operating logic of the entire system. The PCT patent covers this architecture. There are no systems on the market that combine autonomous routing, AI shift scheduling, real-time gamification, and broker attribution in a single integrated platform.

How does Cowboy CRM compare to building a custom internal system?

Building what Cowboy CRM does internally would require: a routing algorithm engineer, a full-stack development team, a DevOps engineer, ongoing maintenance resources, and years of development time. Cowboy CRM represents 9 years of development (2017–2026), three complete rebuild phases, and production validation across 16 enterprise clients and 168,000+ records. The PCT-patented algorithm alone represents years of iteration. Building an equivalent system in-house would cost millions and still lack the production-hardened edge cases that 9 years of real call center data have revealed.

Management & Operations

What does a manager actually do in Cowboy CRM day-to-day?

A manager's primary daily tasks in Cowboy CRM are: reviewing the FTD dashboard and overnight statistics, checking the leaderboard for performance anomalies, reviewing the task escalation queue for items needing intervention, scanning the agent timeline in Calendar for shift coverage issues, and sending a morning floor broadcast in Chat. The shift is pre-built by the AI. Leads are pre-routed. Tasks are pre-assigned. The manager's job is monitoring, coaching, and intervening on exceptions — not operational decisions.

How does a manager know when something needs attention?

The Notifications module routes alerts to managers automatically: task escalations when agents miss deadlines, score drop alerts when an agent's performance falls below threshold, shift coverage alerts when an agent marks absent, and broker alerts when a source's quality drops. Managers do not need to poll the system — the system flags them when action is needed. The goal is that a manager can focus on their best agents' development and strategic planning while the system handles the operational noise.

Can a manager work in Cowboy CRM from a phone or tablet?

The system is responsive and functional on tablet-sized screens. The sidebar collapses on mobile. The most critical manager views — leaderboard, FTD counter, Chat broadcast, and client list — are accessible on mobile. Full configuration and detailed report work is optimized for desktop. Most managers access the system from both desktop (during shift management) and mobile (for monitoring outside office hours).

How does the system handle a scenario where an agent marks absent mid-shift?

When a manager marks an agent as absent in the Shifts module, the system immediately redistributes that agent's assigned active leads to other agents currently on shift, weighted by their current load and performance score. Scheduled callbacks for the absent agent are flagged for reassignment. The manager receives a summary of what was redistributed and to whom. No manual lead-by-lead reassignment is required.

Can different managers see different parts of the floor?

Manager permissions can be scoped by campaign in the Settings → Roles & Permissions configuration. A manager assigned to Campaign A sees agents, leads, and performance data for Campaign A only. Cross-campaign visibility requires admin-level access or explicit permission grants. This is useful in operations where multiple managers are responsible for separate desks or market segments.

How do we monitor the system's health technically?

The Settings → System Health panel shows real-time status of all integrations: database connection and query response time, the Viki Socket.IO engine status and connected client count, email provider delivery rate, SMS gateway delivery rate, and each configured broker API's last successful response. Any degraded component displays in amber; a failed component displays in red with the last successful contact timestamp and the time elapsed since failure. System health alerts are sent to admins via email and SMS immediately on failure detection.

More questions?Contact the Cowboy CRM team through cowyboymanager.com or watch the walkthrough series at the EvolutionCRM YouTube channel. A live demonstration of any specific workflow can be arranged — the system is in active production and can be shown operating with real (anonymized) data.
CEO & Buyers

Why Cowboy CRM

What makes Cowboy CRM categorically different from every other CRM on the market — and why that difference compounds over time into a measurable operational advantage.

9 years in production (2017–2026)
PCT-patented AI routing algorithm
16 enterprise clients, 168k+ records

The Only Autonomous AI Call Center CRM

Cowboy CRM is not a CRM with AI features added. It is an autonomous AI system with a CRM layer built on top. The distinction matters: every other system on the market starts from a contact database and adds routing suggestions, chatbots, or lead scoring overlays. Cowboy CRM starts from the routing algorithm — the decision engine — and everything else (client profiles, finance reports, shift scheduling, gamification) feeds data into and receives commands from that engine.

This architecture is protected under PCT (Patent Cooperation Treaty) international patent. The algorithm was designed in 2017, refined through 3 complete development phases, and validated in production at 16 enterprise call centers across multiple countries. There are no competitors occupying this specific intersection of autonomous routing + shift management + gamification + broker attribution in a single integrated system.

Same algorithm, two productsThe routing AI powering Cowboy CRM is the same engine powering SortExpress.com — the lead generation platform. Clients using both have a seamless pipeline: SortExpress qualifies and scores the lead, Cowboy CRM routes and manages it. No integration gap, no score recalculation, no handoff delay.

Built For This. Not Adapted.

Every major CRM on the market was built for a different use case and adapted for call centers as an aftermarket extension. Salesforce was built for enterprise B2B pipeline management. HubSpot was built for inbound marketing. Zoho was built for SMB general contact management. Each has a call center module or integration that sits awkwardly on top of an architecture designed for something else.

Cowboy CRM was designed from day one for one specific environment: a multi-agent, multi-campaign, multi-broker call center operating in multiple languages at high volume. The shift module, broker module, gamification engine, and communications layer are not plugins — they are native modules sharing the same data model. A shift change immediately affects routing. A broker score drop immediately affects lead allocation. A quiz failure immediately affects an agent's routing priority. This depth of integration is only possible in a system designed for this purpose from the ground up.

9 Years of Production Refinement

Cowboy CRM started development in 2017 under the name Astraq CRM, went through a complete rebuild as EvolutionCRM.ai, and is now deployed as Cowboy CRM at cowyboymanager.com. This is not a new product — it is a mature system that has encountered and solved 9 years of real call center edge cases:

1
Phase 1 (2017–2019): Astraq CRM

Original routing algorithm designed and deployed. Core client management, shift scheduling, and finance attribution built. First enterprise clients onboarded.

2
Phase 2 (2020–2022): EvolutionCRM.ai

Complete architecture rebuild. Gamification engine added. Communications module expanded to three channels. Security hardened to AES-256 with dual-key scheme. SortExpress integration formalized.

3
Phase 3 (2023–2026): Cowboy CRM

PCT patent filed. 14-module system consolidated. 16 enterprise clients in production. 168,000+ client records. Real-time Socket.IO engine (Viki) deployed. Current public release.

14 Deeply Integrated Modules vs. Bolted-On Integrations

CapabilityCowboy CRMGeneric CRM + Integrations
Lead routingNative AI engine, 300ms, 40+ parametersManual assignment or basic round-robin
Shift managementNative module, AI-generated schedulesExternal calendar or HR system, no AI
Broker attributionNative, real-time, 5-dimensionalSpreadsheet or separate platform
GamificationNative, real-time, routing-connectedThird-party add-on, no routing connection
CommunicationsNative, AI-timed, DNC-enforced at APIEmail platform integration, separate SMS
Security auditNative, immutable, every actionPlatform logs only, not action-level
Document managementNative, linked to client records, audit-loggedFile storage integration, no CRM linkage

Development Across 4 Countries

Cowboy CRM was built by development teams across 4 countries: Israel, Ukraine, Bulgaria, and India. This is not a marketing point — it is relevant context for buyers evaluating resilience. The system's architecture was stress-tested by teams operating across different time zones and regulatory environments, which is why it handles multi-currency, multi-language, and multi-country operations natively rather than as configuration options.

What Cowboy CRM Is Not

Cowboy CRM is not a telephony platform — it does not make calls. It routes, schedules, and manages; the calls are made through your existing telephony system via integration. Cowboy CRM is not a marketing automation platform — it sends communications tied to lead events and sequences, not mass marketing campaigns. Cowboy CRM is not a general-purpose CRM adaptable to any industry — it is purpose-built for call centers, and that specificity is its strength. If your operation needs a general CRM for a sales team doing outreach over 6-month cycles, Cowboy CRM is not the right tool. If you run a high-volume outbound call center where routing decisions happen every minute and shift efficiency determines monthly revenue, nothing else on the market does what Cowboy CRM does.

Request a live demonstrationCowboy CRM is in active production with real clients and real data. A demonstration of any specific workflow — routing decisions, shift AI, broker attribution, the gamification floor — can be arranged. Contact the team at cowyboymanager.com.