Workout Window AI Users System

Overview

The AI Users system generates synthetic users to populate workout window teams and squads. This allows:

• Testing the system with users spread across the U.S.
• Ensuring full teams and squads even with low real user adoption
• Real users can naturally replace AI users over time as they join

Database Schema

Migration: 026_add_ai_users.sql

• profiles.is_ai: Boolean flag marking AI users
• workout_window_ai_users: Table storing AI user configuration
  • Region, city, state, timezone
  • Activity type and time window
  • Check-in probability (default 85%)
  • Active status

AI User Generation

Default Configuration

• Regions: 5 U.S. regions with major cities

  • West Coast (25%): LA, SF, Seattle, San Diego, Portland
  • East Coast (30%): NYC, Boston, Philly, DC, Miami
  • Midwest (20%): Chicago, Detroit, Minneapolis, KC, Milwaukee
  • South (15%): Atlanta, Dallas, Houston, Nashville, Austin
  • Mountain West (10%): Denver, Phoenix, Salt Lake City, Las Vegas

• Activity Distribution:

  • Cardio: 30%
  • Strength: 25%
  • Walk: 20%
  • Meditation: 10%
  • Reading: 10%
  • Homework: 3%
  • Other: 2%

• Time Windows: Distributed across day

  • Early morning (6-8am): 15%
  • Morning (8-10am): 10%
  • Lunch (12-2pm): 20%
  • Afternoon (2-4pm): 15%
  • Evening (4-6pm): 25%
  • Late evening (6-8pm): 15%

• Check-in Probability: 85% (configurable)

API Endpoints

Admin Only (Super Admin)

• POST /api/workout-window/admin/ai-users/generate

  • Generate AI users
  • Body: { count: number, config?: Partial<AIGenerationConfig> }
  • Returns: { created, failed, users, errors }

• GET /api/workout-window/admin/ai-users

  • Get all AI users with profiles

• DELETE /api/workout-window/admin/ai-users/:id

  • Delete an AI user (removes all associated data)

• PUT /api/workout-window/admin/ai-users/:id/toggle

  • Toggle active status
  • Body: { isActive: boolean }

• GET /api/workout-window/admin/teams/:date

  • Get team and squad breakdown showing AI vs real users
  • Shows Red/Blue teams, squads, and member breakdown

Admin UI

Accessible via Admin Panel → Workout Window tab.

Features

1. Generate AI Users

   • Specify count (1-1000)
   • Users distributed across U.S. regions
   • Automatically creates:
     • Auth user account
     • Profile with display name
     • Workout window configuration
     • Location data
     • AI user record

2. View AI Users

   • List all AI users
   • Shows: name, location, activity, time window, status
   • Actions: activate/deactivate, delete

3. Team Breakdown

   • View teams and squads for any date
   • Shows AI vs real user counts
   • Color-coded indicators:
     • 🤖 Blue dot = AI user
     • 👤 Green dot = Real user
   • Check-in status (✓ for completed)

4. Squad Breakdown

   • View all 3-person squads
   • Shows team color (Red/Blue)
   • Member breakdown with AI/real indicators

Integration with Chain Matching

AI users are automatically included in chain matching because:

• They have workout_windows entries with is_active = true
• They have workout_window_locations entries with coordinates
• Chain matching queries all active workout windows

No special handling needed - AI users are treated like real users in the matching algorithm.

AI User Replacement

As real users join:

1. Daily chain regeneration includes all active users (AI + real)
2. Real users naturally get matched into chains/squads
3. Over time, real users replace AI users as the system grows
4. AI users can be deactivated as real user density increases

Best Practices

1. Initial Population: Generate 50-100 AI users to ensure full teams
2. Geographic Distribution: AI users are automatically distributed across U.S.
3. Activity Variety: Mix of activity types and time windows
4. Gradual Replacement: As real users join, deactivate AI users in high-density areas
5. Monitoring: Use team breakdown to track AI vs real user ratios

Check-in Behavior

AI users have an 85% check-in probability. This means:

• Most AI users will check in daily
• Some will "miss" check-ins (simulating real behavior)
• Battle chips are earned based on actual check-ins

Future Enhancements

• Automatic Check-ins: Option to auto-check-in AI users based on probability
• Smart Replacement: Automatically deactivate AI users when real users join nearby
• Regional Targeting: Generate AI users in specific regions
• Activity Patterns: More sophisticated activity and time window patterns
• Performance Metrics: Track AI user participation rates