Stage 5 of 6

Intelligence & Services

The layer applications talk to: a unified profile, a fast query engine, AI-driven insight, and the operational plumbing that keeps everything running and billed correctly.

👤 Identity & Profile Service

How to Consume

Cxos.Profile.Api · Query

GET /v1/profile/cust_004821
Authorization: Bearer <service-token>

200 OK
{ "customer_key": "cust_004821", "lifecycle_stage": "active", "ltv_band": "gold" }
-- fields the caller isn't entitled to are omitted, not nulled

📊 Query & Analytics Engine

DataFusion (SQL Engine)

How to Consume

Cxos.Intelligence.Api · Query

POST /v1/query
{ "metric": "net_revenue", "dimensions": { "region": "APAC" }, "time_grain": "week" }

200 OK
{ "net_revenue": 51230.40, "cache_hit": true }

🧠 AI & Insights

How to Consume

Cxos.Intelligence.Api · Both

GET /v1/propensity/cust_004821
{ "churn_risk": 0.78, "conversion_propensity": 0.12, "model_version": "churn_risk_v4" }

POST /v1/copilot/ask { "question": "which APAC customers are at risk of churning?" }

⚙ Operational Services

How to Consume

Cxos.Operations.Api · Both

GET /v1/workflows/wf_88213
{ "type": "schema_change_approval", "state": "pending_approver" }

GET /v1/usage/tenant_northwind?period=2026-07
{ "ingestion_events": 812400000, "estimated_cost_usd": 6240.18 }

🔌 Platform Connectors

What each AI & analytics capability connects out to, beyond the lakehouse it reads from.

Connector PackageConsumer
Cxos.Connectors.Salesforce / .Zendesk / .HubSpot (read path) Unified Customer Profile · Identity Graph (identity resolution inputs)
Azure OpenAI Service client Natural Language Query · AI Copilot
Azure Machine Learning client Predictive Models · Propensity Scores · Anomaly Detection
DataFusion Iceberg connector Query Optimizer · Semantic Layer
Real World Example

You query: "Show me users who viewed product X, added to cart, but didn't purchase in the last 7 days." The engine returns results in milliseconds. An AI model predicts who is likely to convert.