AI Usage Statistics (Super Admin)

Track AI usage across your organization with comprehensive statistics showing total calls, user vs. company API usage, and monthly trends.

Super Admin Only

This feature is only available to Super Admin users. Organization Admins and regular users cannot access AI Usage Statistics.

Overview

What Are AI Usage Statistics?

AI Usage Statistics provide visibility into how your organization uses AI features:

• Total API Calls - All AI requests across all users
• User-Provided APIs - Calls using users' personal API keys
• Global API Calls - Calls using company-provided global API key
• Monthly Usage - Current billing period activity

Why Track AI Usage?

Cost Management:

• See how much company is spending on AI (Global API Calls)
• Understand user vs. company cost distribution
• Make informed decisions about AI strategy

Usage Patterns:

• Identify power users (high API usage)
• See adoption rate across organization
• Detect unusual usage patterns

Billing Strategy:

• Determine if global AI is cost-effective
• Decide whether to encourage personal API keys
• Plan budget for AI features

Accessing AI Usage Statistics

Navigation

1. Login as Super Admin

   • Must have super_admin role
   • Organization Admins cannot access this feature

2. Go to Super Admin Dashboard

   • Click "Super Admin" in sidebar
   • Scroll to "AI Usage Statistics" section

3. View Dashboard

   • Statistics display automatically
   • No "Show" button required (always visible)

Dashboard Location

Super Admin Dashboard
├── Organization Management
├── User Management
├── Global AI Settings
└── AI Usage Statistics ← HERE
    ├── Total API Calls (blue card)
    ├── User-Provided APIs (purple card)
    ├── Global API Calls (green card)
    └── This Month (orange card)

Understanding the Metrics

Metric Cards

The dashboard displays four gradient cards with key statistics:

1. Total API Calls (Blue)

┌────────────────────────────┐
│ Total API Calls            │
│                            │
│     1,247                  │
│                            │
│ All organizations          │
└────────────────────────────┘

What it shows: Total number of AI API calls across all organizations

Includes:

• ✅ User API calls (personal keys)
• ✅ Global API calls (company key)
• ✅ All organizations (platform-wide)
• ✅ All time (since system launch)

Use Case: Overall platform AI activity

2. User-Provided APIs (Purple)

┌────────────────────────────┐
│ User-Provided APIs         │
│                            │
│      387                   │
│                            │
│ Users' own keys            │
└────────────────────────────┘

What it shows: API calls using users' personal API keys

Includes:

• ✅ Users who configured personal AI settings
• ✅ Calls using user's API key (user pays)
• ❌ NOT using company global key

Use Case: See how many users prefer their own AI

3. Global API Calls (Green)

┌────────────────────────────┐
│ Global API Calls           │
│                            │
│      860                   │
│                            │
│ Company-provided key       │
└────────────────────────────┘

What it shows: API calls using company's global AI settings

Includes:

• ✅ Users without personal AI settings
• ✅ Calls using global API key (company pays)
• ❌ NOT using user personal keys

Use Case: See company's AI spend

Cost Calculation

Global API Calls × Provider Cost = Your Monthly Bill

Example:

• 860 calls × $0.001 (Gemini) = $0.86/month
• 860 calls × $0.06 (GPT-4) = $51.60/month

4. This Month (Orange)

┌────────────────────────────┐
│ This Month                 │
│                            │
│      214                   │
│                            │
│ Current billing period     │
└────────────────────────────┘

What it shows: Total API calls in current calendar month

Includes:

• ✅ User API calls (this month)
• ✅ Global API calls (this month)
• ❌ Resets on 1st of month

Use Case: Track monthly trends and budget

Relationship Between Metrics

Formula:

Total Calls = User API Calls + Global API Calls

Example:

Total: 1,247 calls
├── User API: 387 calls (31%)
└── Global API: 860 calls (69%)

Interpretation:
- 69% of users rely on company-provided AI
- 31% of users bring their own API keys

Interpreting the Statistics

Scenario 1: High Global API Calls

Numbers:

Total: 1,000
User API: 100 (10%)
Global API: 900 (90%)
This Month: 300

Interpretation:

• 90% of users rely on company-provided global AI
• Only 10% use personal API keys
• Company paying for most AI usage

Action:

• ✅ If budget allows: Continue providing global AI (good user experience)
• ⚠️ If costs too high: Encourage users to configure personal API keys
• 💡 Consider: Switch global AI to cheaper provider (Gemini vs GPT-4)

Scenario 2: High User API Calls

Numbers:

Total: 1,000
User API: 800 (80%)
Global API: 200 (20%)
This Month: 250

Interpretation:

• 80% of users prefer their own API keys
• Only 20% use company-provided AI
• Low company AI costs

Action:

• ✅ Keep global as fallback: Provides safety net for new users
• ✅ Low cost: Company spends minimal on AI
• 💡 Consider: Global AI is working as intended (optional fallback)

Scenario 3: Balanced Mix

Numbers:

Total: 1,000
User API: 500 (50%)
Global API: 500 (50%)
This Month: 300

Interpretation:

• 50/50 split between user and company AI
• Balanced approach working well
• Moderate company costs

Action:

• ✅ Ideal setup: Flexibility for users, reasonable company cost
• ✅ Maintain status quo: Both options available

Scenario 4: Growing Monthly Usage

Numbers:

Total: 5,000 (all time)
This Month: 1,200
Last Month: 800 (estimated)

Interpretation:

• AI usage growing 50% month-over-month
• Users adopting AI features more
• Monthly costs increasing

Action:

• 📊 Monitor trend: Is growth sustainable?
• 💰 Budget planning: Project next 3-6 months
• 🔄 Optimize: Consider cheaper model for global AI

Making Decisions Based on Statistics

Question 1: Should We Provide Global AI?

Check: Global API Calls count

High Global Usage (>70%):

• ✅ Yes, provide global AI - Users rely on it heavily
• Benefits: Good user experience, easier onboarding
• Trade-off: Company pays for AI

Low Global Usage (less than 30%):

• 🤔 Maybe not needed - Most users have personal keys
• Consider: Disable global, require personal keys
• Benefit: Zero company AI cost

Balanced (30-70%):

• ✅ Keep global as fallback - Best of both worlds

Question 2: Which Provider Should We Use?

Check: Global API Calls count × Cost per call

For High Volume (>5,000 calls/month):

• Use cheapest provider: Google Gemini (~$5-10/month)
• Avoid: GPT-4 (~$300/month)

For Low Volume (under 1,000 calls/month):

• Can afford better quality: ChatGPT or Claude
• Cost acceptable even for premium models

For Balanced:

• Use Gemini 2.0 Flash (fast + affordable)

Question 3: Should We Encourage Personal API Keys?

Check: User API Calls percentage

High User API (>60%):

• ✅ Already happening - Users prefer personal keys
• Keep documenting: How to configure personal AI

Low User API (under 30%):

• 💡 Consider encouraging:
  • Send email: "Configure personal AI for unlimited usage"
  • Add banner in app: "Want more AI? Add your own API key"
  • Benefits: Reduce company costs

Reasons Users Prefer Personal Keys:

• ✅ Unlimited usage (within their API limits)
• ✅ Choose preferred provider
• ✅ Full control over costs
• ✅ Faster rate limits

Question 4: Is Our AI Budget Sufficient?

Check: This Month calls × Cost

Calculate Monthly Budget Need:

Example:
This Month: 1,200 global API calls
Provider: ChatGPT GPT-3.5-turbo
Cost: ~$0.002 per call

Monthly cost: 1,200 × $0.002 = $2.40/month ✅ Affordable

vs.

Provider: ChatGPT GPT-4
Cost: ~$0.06 per call

Monthly cost: 1,200 × $0.06 = $72/month ⚠️ Review budget

Decision Matrix:

Monthly Calls	Gemini Cost	GPT-3.5 Cost	GPT-4 Cost	Recommendation
500	~$0.50	~$1	~$30	Any model OK
1,000	~$1	~$2	~$60	GPT-4 if budget allows
5,000	~$5	~$10	~$300	Use Gemini or GPT-3.5
10,000	~$10	~$20	~$600	Use Gemini, encourage user keys

Best Practices

Regular Monitoring

Weekly Check:

• View dashboard weekly
• Note any sudden spikes
• Identify anomalies early

Monthly Review:

• Compare month-over-month growth
• Adjust global AI provider if needed
• Update budget projections

Quarterly Analysis:

• Review user adoption trends
• Calculate total AI spend
• Make strategic decisions

Setting Up Alerts

On Provider Dashboard:

Most AI providers allow billing alerts:

Example (OpenAI):

Settings → Billing → Usage Limits
├── Set monthly budget: $50
├── Email alert at 75%: $37.50
└── Hard limit at 100%: $50

Example (Google Cloud - Gemini):

Billing → Budgets & Alerts
├── Create budget: $20/month
├── Alert at 50%: $10
├── Alert at 90%: $18
└── Alert at 100%: $20

Educating Users

Communicate Statistics:

Send monthly email to organization:

Subject: NirmIQ AI Usage - January 2024

Team,

Our AI feature usage this month:
• Total calls: 1,200 (+15% from December)
• 70% using company-provided AI (free for you!)
• 30% using personal API keys

Want unlimited AI usage?
Configure your personal AI settings in Settings page.
Personal API keys give you:
  ✓ Unlimited usage
  ✓ Choice of provider
  ✓ Full control

Questions? Check our AI Integration Guide.

Thanks,
IT Team

Cost Optimization Strategies

1. Start with Free Tier

• Gemini: 15 requests/minute free
• Test with free tier before committing budget

2. Hybrid Approach

• Power users: Encourage personal API keys
• Casual users: Provide global AI
• Result: Lower company costs

3. Right-Size Model

• Don't use GPT-4 if GPT-3.5 suffices
• Requirements rewriting doesn't need most advanced model
• Test quality vs. cost trade-off

4. Monitor and Adjust

• Review monthly: Is current provider/model optimal?
• A/B test: Try different providers for global AI
• Gather feedback: Ask users about AI quality

Troubleshooting

Statistics Show Zero

Issue: All metrics show 0

Possible Causes:

1. No AI usage yet - Organization just started
2. Global AI not configured - No users have AI access
3. Database error - Statistics not loading

Solution:

• Check if Global AI Settings configured
• Verify users are using AI features
• Test AI analysis on a requirement

User API vs Global API Don't Add Up

Issue: User API + Global API ≠ Total

Possible Causes:

1. Database sync delay - Statistics refresh every request
2. Deleted usage records - Rare, shouldn't happen
3. Multi-organization - Total includes all orgs

Solution:

• Refresh page (F5)
• Check again in few minutes
• Contact support if persists

This Month Seems Wrong

Issue: "This Month" shows unexpected number

Possible Causes:

1. Timezone difference - Month boundary varies by timezone
2. Cached data - Browser caching old statistics
3. Recent usage not counted - Background job delay

Solution:

• Hard refresh (Ctrl+F5 or Cmd+Shift+R)
• Check timezone settings
• Wait 5 minutes and refresh

Sudden Spike in Usage

Issue: Calls jumped from 100/month to 5,000/month

Investigation:

1. Check Recent Activity:

   • Did someone bulk import requirements?
   • Did team start new large project?
   • Is AI being used incorrectly?

2. Identify Heavy Users:

   • (Future feature: per-user usage breakdown)
   • For now: Check recent requirement edits

3. Verify Legitimate:

   • Spike might be normal (new project launch)
   • Or could indicate API key leak/abuse

Actions:

• If legitimate: Adjust budget, switch to cheaper model
• If suspicious: Rotate API keys, investigate users
• Set up provider-side rate limits

Future Enhancements

Coming soon:

• 📊 Per-User Usage Breakdown - See which users use AI most
• 📈 Historical Charts - Visualize usage trends over time
• 🔔 Usage Alerts - Email when thresholds exceeded
• 💰 Cost Tracking - Automatic cost calculation per provider
• 📅 Date Range Filters - View usage for custom time periods
• 📤 Export Statistics - Download CSV for analysis
• 🎯 Per-Organization Stats - Break down by organization (Super Admin)

Related Documentation

• Global AI Settings - Configure company-wide AI
• AI Integration Guide - User-facing AI documentation
• User Management - Managing user roles and permissions

---

Questions? Contact support or visit our Training Portal