AI Engine Module

Ask your association data questions in natural language and get intelligent answers.

What Can You Do?

🟢 Easy — Click Around

Chat Assistant — Ask questions in plain English. The AI knows your data and answers using real numbers.

AI Insights — Get automatic analysis of trends across your association.

Semantic Search — Find documents by meaning, not just keywords.

Churn Prediction — AI identifies members at risk of leaving.

Anomaly Detection — Catches unusual financial transactions or attendance patterns.

Document Generation — AI creates polished reports, minutes, and letters.

Try it now:

1. Click AI Engine in the sidebar
2. Type: "How many active members do we have?"
3. The AI answers with your actual data
4. Try: "What events are coming up this month?"
5. Try: "Show me the revenue breakdown"

🔵 Advanced — API / Code

API Endpoints (11)

Method	Endpoint	Description	Auth
POST	/ai/chat	Chat with AI	Everyone
GET	/ai/health	AI service health	Everyone
GET	/ai/insights	Auto-generated insights	Staff+
POST	/ai/embeddings/search	Semantic search	Staff+
GET	/ai/models	Available AI models	Staff+
POST	/ai/predict/churn/{member_id}	Churn prediction	Staff+
POST	/ai/predict/anomalies	Anomaly detection	Staff+
POST	/ai/generate/document	AI document generation	Staff+
GET	/ai/conversations/{session_id}	Conversation history	Member
DELETE	/ai/conversations/{session_id}	Delete conversation	Member
GET	/ai/embeddings/stats	Embedding statistics	Staff+

LLM Configuration

• Provider: OpenRouter (primary), Groq (fallback)
• Primary Model: meta-llama/llama-3.1-8b-instruct
• Fallback Models: google/gemma-4-31b-it:free → llama-3.1-8b-instant
• Embeddings: pgvector with 384-dimensional vectors

Example: Chat

curl -X POST https://ams.14.jugaar.ai/api/v1/ai/chat \
  -H "Authorization: Bearer $TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"message": "How many active members do we have?"}'

Expected response (200):

{
  "response": "You have 57 active members out of 78 total members...",
  "session_id": "uuid",
  "model_used": "meta-llama/llama-3.1-8b-instruct",
  "sources": ["members_table", "member_stats"]
}

Example: Churn Prediction

curl -X POST https://ams.14.jugaar.ai/api/v1/ai/predict/churn/{member_id} \
  -H "Authorization: Bearer $TOKEN"

Returns: Risk score (0-1), contributing factors, recommended actions.

Example: Anomaly Detection

curl -X POST https://ams.14.jugaar.ai/api/v1/ai/predict/anomalies \
  -H "Authorization: Bearer $TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"module": "finances", "lookback_days": 30}'

---

Related

• Testing: AI Engine
• Members: Churn Prediction
• Analytics