Skip to main content

AI Integration

NirmIQ integrates with multiple AI providers to enhance requirement analysis, validation, and quality improvement.

Supported AI Providers​

NirmIQ supports all major commercial AI providers — including Google Gemini, OpenAI (ChatGPT), Anthropic (Claude), xAI (Grok), and DeepSeek. The model dropdown inside Settings shows the models currently available for your chosen provider.

Choosing a model
  • For most users, the default model offered by your selected provider is the right starting point.
  • Faster, lower-cost variants are good for high-volume tasks like document import; flagship models give the best quality on complex reasoning tasks.
  • You can switch providers and models at any time without losing data.

How AI Settings Work​

NirmIQ supports two ways to power AI features:

  1. Bring Your Own Key (BYOK) — configure your personal API key from your preferred provider in Settings → AI Integration. You pay the provider directly, and your settings always take precedence.
  2. Organisation-managed AI — your Super Admin configures a shared API key for the whole organisation. Used as a fallback for any user who hasn't set up their own.

If neither is configured, you'll see "AI not configured" and be prompted to set one up.

User AI Settings​

Configuring Your Personal AI​

  1. Navigate to Settings (gear icon in header)
  2. Scroll to AI Integration section
  3. Select your preferred AI Provider
  4. Choose the Model (auto-populated based on provider)
  5. Enter your API Key
  6. Check Enable AI features
  7. Click Save AI Settings
tip

The AI Settings panel is located in Project Settings > AI Configuration. Look for the gear icon in the top navigation bar.

Getting API Keys​

ProviderAPI Key Location
Google GeminiGoogle AI Studio
ChatGPT (OpenAI)OpenAI Platform
Claude (Anthropic)Anthropic Console
Grok (xAI)xAI Platform
DeepSeekDeepSeek Platform

Leaving Settings Blank​

If you don't configure personal AI settings:

  • ✅ System uses global AI settings (if configured by Super Admin)
  • ✅ You still get AI features
  • ⚠️ Usage counted against company quota
  • 💡 Best for: Teams with centralized AI management

If you configure personal AI settings:

  • ✅ Your settings always take priority
  • ✅ Unlimited usage with your own API key
  • ✅ Choose your preferred provider
  • 💰 Costs charged to your API account
  • 💡 Best for: Power users, specific provider preference, unlimited usage needs

Super Admin: Global AI Settings​

Overview​

Super Admins can configure global AI settings that apply to all users who haven't set up personal AI configuration.

Access: Super Admin Dashboard → Global AI Settings

Configuring Global Settings​

  1. Click Show Settings button
  2. Select AI Provider for company-wide use
  3. Enter Model name
  4. Add company API Key
  5. Check Enable global AI settings
  6. Click Save Global AI Settings
Company-Wide Control

Global settings are perfect for:

  • Providing AI features to all team members
  • Centralized billing and usage tracking
  • Standardizing on one AI provider
  • Controlling costs through company API limits
User Settings Override

Remember: User personal settings always override global settings. This gives users flexibility while providing a fallback option.

AI Usage Statistics​

For Super Admins​

Track AI usage across your organization:

Dashboard Location: Super Admin Dashboard → AI Usage Statistics

Metrics Available:

  • Total API Calls - All AI requests across all organizations
  • User-Provided APIs - Calls using users' personal API keys
  • Global API Calls - Calls using company-provided global API key
  • This Month - Current billing period usage
info

The AI Usage Statistics dashboard shows a breakdown of API calls by source (user keys vs global keys) with monthly trends.

Understanding the Metrics​

Total Calls = User API Calls + Global API Calls

Use Case: Determine if you should provide global AI or encourage users to bring their own keys.

Example:

  • Total: 1,000 calls
  • User APIs: 800 calls
  • Global APIs: 200 calls
  • Insight: 80% of users prefer their own API keys

AI Features in NirmIQ​

Requirement Analysis​

What it does: Analyzes and improves requirement text using AI.

How to use:

  1. Open any requirement in the List tab
  2. Click the AI Analyze button (brain icon)
  3. Review the AI-suggested improvement
  4. Accept or modify the suggestion

What you get back: A cleaner, more standardised version of your requirement that preserves the original intent. Review the suggestion before accepting.

Document Import (AI-Powered)​

What it does: Automatically extracts and structures requirements from Word documents.

How to use:

  1. Click Import button in Requirements workspace
  2. Select your .docx file
  3. (Optional) Upload methodology document
  4. AI extracts requirements and creates hierarchy

AI Processing:

  • Identifies requirement statements
  • Assigns hierarchy levels
  • Generates requirement IDs
  • Creates parent-child relationships

Error Messages​

NirmIQ provides clear, actionable error messages when AI operations fail:

Common Errors and Solutions​

Error MessageMeaningSolution
"Invalid API key for [Provider]"API key is incorrect or expiredCheck your API key in Settings
"Rate limit exceeded"Too many requestsWait a moment, or upgrade API plan
"Billing issue with [Provider]"Account has no creditsAdd credits or payment method
"[Provider] experiencing technical difficulties"Provider server errorTry again later
"AI not configured"No settings foundConfigure AI in Settings page

Error Handling Features​

✅ User-Friendly Messages - Clear, actionable guidance instead of technical jargon

✅ Provider-Specific Help - Tailored advice for each AI provider

✅ Automatic Retries - Some errors trigger automatic retry logic

✅ Detailed Logging - Technical details logged for support (when needed)

Best Practices​

For Individual Users​

  1. Start with Your Own API Key

    • More control over costs
    • Choose your preferred provider
    • Unlimited usage (within your API limits)

  2. Test Different Providers

    • Each AI has different strengths
    • Try multiple providers for your use case
    • Switch providers anytime

  3. Monitor Your API Usage

    • Check your provider's dashboard
    • Set up billing alerts
    • Understand your costs

For Organizations​

  1. Provide Global Settings as Fallback

    • Configure global AI for all users
    • Users without personal keys automatically get access
    • You control the budget through API limits

  2. Track Usage Patterns

    • Review AI Usage Statistics
    • Understand user vs company API usage
    • Make informed decisions on AI strategy

  3. Educate Your Team

    • Share this documentation
    • Explain priority system
    • Encourage users to configure personal settings if needed

Security & Privacy​

API Key Storage​

  • ✅ API keys stored encrypted in database
  • ✅ Displayed as masked passwords in UI
  • ✅ Never logged or exposed in error messages
  • ✅ Transmitted over HTTPS only

Data Handling​

  • Your requirements are sent to AI providers for analysis
  • Review your chosen provider's data policy
  • Consider using company-approved providers
  • Data is not stored by AI providers (in most cases)

Access Control​

  • User Settings: User can view/edit only their own
  • Global Settings: Super Admin only
  • Usage Stats: Super Admin only
  • Row-Level Security: Enforced at database level

Troubleshooting​

AI Button Not Working​

Check:

  1. Is AI enabled? Settings → AI Integration → "Enable AI features"
  2. Do you have valid API key configured?
  3. Is the API key active and has credits?

AI Analysis Returns Error​

Try:

  1. Check error message for specific issue
  2. Verify API key is correct
  3. Test API key directly with provider
  4. Check provider's status page
  5. Contact support if issue persists

Can't Save AI Settings​

Possible Causes:

  1. Missing required fields (provider, model, API key)
  2. Invalid API key format
  3. Network connectivity issue
  4. Permission issue (contact admin)

FAQ​

Q: Which AI provider should I use?​

A: It depends on your needs:

  • Budget-conscious: Google Gemini (free tier available) or GPT-4o-mini
  • Best quality: GPT-4o or Claude 3.5 Sonnet
  • Fastest: Google Gemini 2.0 Flash
  • Complex reasoning: Claude 3 Opus or OpenAI o1
  • Specific requirements: Test different providers

Q: What happens if I change providers?​

A: Your settings are saved immediately. Next AI operation uses new provider.

Q: Can I use different providers for different projects?​

A: AI settings are user-level today — all the projects you participate in use the same AI configuration that you have set on your own profile. There is no per-project AI provider override. (If you need project-scoped AI — for example, different providers for different clients — please email info@nirmiq.com; it is on the roadmap.)

Q: Does NirmIQ charge for AI usage?​

A: There are two ways AI calls are billed:

  1. Bring Your Own Key (BYOK) — you save your own API key in Settings → AI. You pay your AI provider directly; NirmIQ does not see those bills. BYOK users are not subject to the AI add-on check.
  2. NirmIQ-managed AI — your organisation enables the AI add-on and NirmIQ uses its shared keys. Usage is metered against your tier's monthly AI allowance and the add-on's seat-based pricing.

If you have configured a personal API key, that key is used. Otherwise, NirmIQ falls back to the organisation's NirmIQ-managed AI — which is only available if the AI add-on is enabled for your organisation.

Q: What if both my personal key and the organisation's managed AI are available?​

A: Your personal key wins. The resolver checks your Settings → AI first; if a key is present and active, it is used. Only when you have no personal key does NirmIQ fall back to the managed AI. This means BYOK users are unaffected by changes to the add-on.

Q: Can I disable AI for certain users?​

A: Yes. Admin can toggle "AI Permission" for each user in User Management.

Next Steps​

Questions? Contact support or visit our Training Portal