Strykr AI Assistant — Product Specification
Version: 1.0 Date: March 2026 Status: Draft
Overview
The Strykr AI Assistant is a two-part intelligent system: a Help & Support system that resolves user issues instantly or escalates them with full context, and a Sports Intelligence engine that helps users find edges through deep analysis of matches, players, and markets.
Part 1: Help & Support
Vision
A support system that resolves 80% of user issues instantly through AI-powered understanding, and seamlessly escalates the rest with full context — no repetition, no lost information.
Core Flow
User opens Help & Support
|
v
Describes issue (text)
|
v (optional)
Attaches evidence (screenshot / video)
|
v
AI Attempts Resolution
|
+----+----+
| |
Resolved Unresolved
| |
v v
Done Creates Ticket
|
v
Auto-assigns priority
Auto-routes to team
|
v
User tracks in Active Tickets
1.1 Smart Resolution
Before creating a ticket, the AI assistant attempts to resolve the issue using its understanding of the platform and the user's account state.
What it can resolve instantly:
- "How do I..." questions — guided walkthroughs
- Balance/credit confusion — explains Take, Credit Limit, Available Credit with the user's actual numbers
- Bet status questions — "why hasn't my bet settled?" with real-time order status check
- Feature discovery — "can I cancel a bet?" with step-by-step instructions
- Settlement cycle questions — current period status, when grace starts
- Known issues — acknowledges with workaround if available
- Commission questions — explains how commission works with user's actual commission data
What it escalates:
- Stuck/failed bets that should have settled
- Balance discrepancies that don't match expected calculations
- Provider errors (Betfair, Bifrost, Pinnacle)
- Access/permission issues
- Bugs with reproducible steps
- Feature requests
1.2 Ticket Creation
When the AI cannot resolve an issue, it automatically creates a ticket:
AI-Generated Fields:
- Title: Concise summary extracted from conversation
- Description: Structured summary of the issue, steps to reproduce, expected vs actual behavior
- Priority: Auto-assigned based on impact
- P0 — Critical: Financial impact, funds stuck, incorrect settlement
- P1 — High: Cannot place bets, settlement blocked, login issues
- P2 — Medium: UI bugs, display errors, slow performance
- P3 — Low: Questions, suggestions, cosmetic issues
- Category: Billing, Betting, Technical, Account, Feature Request
- Assigned To: Auto-routed based on category
- Attachments: Screenshots and video recordings from the conversation
What the user provides:
- Natural language description of the issue
- Optional screenshot (upload or paste)
- Optional screen recording (for complex interaction bugs)
What the AI adds:
- Platform context at time of report (user's balance, open bets, browser, device)
- Conversation transcript leading to ticket creation
- Suggested category and priority (user can override)
1.3 Evidence Capture
Users can attach rich evidence to help the team understand and reproduce issues:
Screenshots
- Direct upload from device
- Paste from clipboard
Video Recording
- In-app screen recording (capture interactions that led to the bug)
- Upload from device
- Maximum duration: 2 minutes
- Auto-compressed for storage
1.4 Active Tickets
Users can track all their open tickets:
- Ticket status: Open, In Progress, Awaiting Response, Resolved
- Team member assigned
- Last update timestamp
- Add comments or additional evidence
- Reopen resolved tickets if issue recurs
1.5 Past Tickets
Complete history of resolved tickets:
- Searchable by date, category, status
- Resolution details
- Time to resolution
1.6 Agent & Admin Ticket Views
Agent View:
- See tickets from their downline players
- Add context or comments
- Escalate on behalf of player
Admin View:
- All tickets across the platform
- Filter by priority, category, status, assigned team member
- Bulk actions (reassign, reprioritize)
- SLA tracking (time to first response, time to resolution)
- Analytics: common issues, resolution rates, average response time
Part 2: Strykr Intelligence (Chat)
Vision
A context-aware sports analyst that combines live match data, historical statistics, player intelligence, and real-time news to surface actionable trading edges — before, during, and after matches.
2.1 Pre-Match Analysis
The assistant ingests and synthesizes multiple data sources to provide pre-match intelligence:
Team & Match Context
- Head-to-head records between teams
- Recent form (last 5/10 matches)
- Venue statistics (ground averages, pitch behavior, toss advantage)
- Weather and conditions impact
- Team composition changes (injuries, squad announcements, playing XI)
Player Intelligence
- Individual player form and recent performance
- Player strengths and weaknesses (e.g., "plays hook shot best", "struggles against short-pitched bowling")
- Player vs player matchup data (e.g., "Kohli averages 12 against Bumrah in T20s")
- Bowler effectiveness against batting styles (e.g., "struggles to bowl left-handed batsmen")
- Player performance by venue, phase of innings, and conditions
- Strike rates by phase (powerplay, middle overs, death)
- Bowling economy by phase and match situation
Odds Intelligence
- Current market odds with movement trends
- Opening odds vs current odds (line movement analysis)
- Volume and liquidity analysis
- Value identification: where market odds diverge from model-estimated probability
- Historical odds accuracy for similar matchups
News & Sentiment
- Real-time sports news aggregation from multiple sources
- Social media sentiment (X/Twitter, sports forums)
- Injury updates, team news leaks
- Expert predictions and pundit analysis
- Toss result and its historical impact on outcomes
2.2 In-Play Analysis (Live Match Intelligence)
During a live match, the assistant processes real-time data and surfaces trading opportunities when asked.
Ball-by-Ball Intelligence
- Live ball-by-ball data feed ingestion
- Run rate tracking: current, required, projected
- Momentum detection: identifying shifts before the market reacts
- Key moments: wickets, boundaries, milestones, partnerships — with odds impact analysis
- Phase analysis: where the match is vs expected trajectory
Live Player Matchup Analysis
- Current batter vs current bowler historical data
- Batter's form in this innings vs career averages
- Bowler's spell analysis (economy, dot ball %, wicket probability)
- Scoring zones and fielding placement intelligence
Predictive Context
- Win probability shifts with explanation
- Projected totals based on current run rate, wickets in hand, and conditions
- Phase-specific projections (e.g., "at 120/2 after 15 overs, teams batting first here average 185")
- Death bowling matchup preview based on batting order remaining
2.3 Post-Match & Bet Analysis
Personal Performance Review
- Analysis of user's bet history (win rate, ROI, patterns)
- Edge identification: which sports, markets, or bet types the user performs best on
- Leak detection: recurring mistakes (e.g., "you consistently overvalue home teams in T20s")
- Suggested adjustments based on historical patterns
Market Review
- Post-match odds review: where was value actually found?
- Key turning points mapped to odds movements
2.4 User Context Awareness
The assistant knows who it's talking to:
- Current balance and exposure
- Open bets and their status
- Risk profile based on betting history
- Role-appropriate responses (player vs agent vs admin)
- Personalized stake suggestions based on bankroll
2.5 Data Sources
| Source | Type | Update Frequency |
|---|---|---|
| Ball-by-ball feed | Live match data | Real-time (every ball) |
| Exchange odds | Market data | Real-time |
| Betfair/Pinnacle | Historical odds | Pre-match + in-play |
| News aggregators | Text | Every 5 minutes |
| X/Twitter | Social sentiment | Real-time |
| Player stats DB | Historical | Daily refresh |
| Head-to-head records | Historical | Daily refresh |
| Venue/pitch data | Historical | Per match |
| Weather APIs | Conditions | Hourly |
| Team announcements | Squad/XI | As available |
What Gets Removed
The current "Command" tab is deprecated. Its useful capabilities are redistributed:
| Current Command Feature | Where It Goes |
|---|---|
| Place bet via chat | Removed — users bet through the UI |
| Check balance via chat | Removed — visible in wallet |
| View matches via chat | Removed — browse in Sports tab |
| Navigate to page | Removed — direct navigation |
| Bet history analysis | Moved to Strykr Intelligence |
| User context awareness | Moved to Strykr Intelligence |
| Agent balance adjustments | Removed — use Edit Credit Limit in Downline |
Information Architecture
AI Assistant Panel
|
+-- Tab 1: Help & Support
| |-- Smart resolution chat
| |-- Create ticket (with attachments)
| |-- Active Tickets (list + detail)
| +-- Past Tickets (searchable history)
|
+-- Tab 2: Strykr Intelligence
|-- Pre-match analysis
|-- In-play live analysis
|-- Bet history review
|-- Personalized insights
+-- Conversational (ask anything about sports/markets)
Success Metrics
Help & Support:
- AI resolution rate: % of issues resolved without ticket creation (target: 80%)
- Time to resolution: Average time from ticket creation to resolution
- First response time: Average time for team to acknowledge a ticket
- Ticket volume trend over time (should decrease as AI improves)
- User satisfaction: Post-resolution rating
Strykr Intelligence:
- Engagement: % of active users who interact with the analyst weekly
- Quality: User rating on analysis accuracy (thumbs up/down)
- Edge conversion: % of AI-surfaced opportunities that were acted on
- In-play usage: Sessions where in-play analysis was accessed during live matches
Phased Rollout
Phase 1: Help & Support
- AI-powered issue resolution chat
- Ticket creation with screenshot/video upload
- Active and past ticket views
- Admin ticket management
Phase 2: Enhanced Strykr Intelligence
- User context integration (balance, bets, role)
- Bet history analysis and personal performance review
- Pre-match analysis with player intelligence data
Phase 3: In-Play Intelligence
- Ball-by-ball data feed integration
- Live player matchup analysis
- Real-time odds context in analysis
- Predictive modeling and win probability
Phase 4: Deep Player Intelligence
- Player strengths/weaknesses database
- Bowler vs batter matchup matrix
- Venue and conditions modeling
- Historical odds accuracy tracking