Vehicle Owner Companion
Outcome
Answer a vehicle owner's questions about their exact VIN configuration, warning lights, and service needs using the owner manual, build configuration, and connected telemetry, with dealer confirmation before any booking or remote command.
| Portfolio attribute | Value |
|---|---|
| Primary persona | Vehicle owner |
| Primary aggregate | VehicleOwnership |
| Cosmos DB containers | 2 |
| Grounding data | manuals, VIN configuration, telemetry, service history |
| Development estimate | USD 60-220 monthly dev estimate |
| Runtime options | Python FastAPI and C# ASP.NET Core |
Productivity hypothesis
These measures reflect the workflow, evidence, decision boundary, and industry risk for this use case.
| Impact type | Measure | Conservative internal assumption | Customer measurement |
|---|---|---|---|
| Productivity | Response cycle time | 10-20% reduction | (baseline median response minutes - assisted median response minutes) / baseline median response minutes * 100 |
| Quality or risk | Evidence discovery time | 15-25% reduction | (baseline median evidence-search minutes - assisted median evidence-search minutes) / baseline median evidence-search minutes * 100 |
Important: Conservative internal pilot assumptions only; not public benchmarks, commitments, or expected customer results. Validate against a comparable baseline or matched control cohort with mandatory human review.
Rationale: Vehicle-owner support can reduce manual lookup across manuals and service data without replacing professional diagnosis.
Pilot guardrail: Keep guidance within the evidence for the specific VIN, configuration, telemetry, and service history.
Customer validation plan
- Baseline the current workflow with representative users and tasks.
- Run the accelerator on the same task mix with human review retained.
- Measure response cycle time and evidence discovery time with the documented formulas.
- Replace internal assumptions with customer-observed results before a business case.
Why Azure Cosmos DB
This AI solution needs an operational data plane, not only a vector index:
- Hybrid vector and full-text retrieval combines semantic recall with exact identifiers, codes, names, and domain language.
- Hierarchical partition keys and metadata predicates keep tenant and domain boundaries in the normal query path.
- Operational JSON, source chunks, workflow and agent state, feedback, and audit records remain together for low-latency reads and governed writes.
- Change feed reacts to source and operational updates without polling.
- Autoscale throughput, configurable consistency, and multi-region distribution support variable demand and enterprise availability.
Domain workflow

| Stage | Actor | Input → output | Approval |
|---|---|---|---|
| 1. Ask a vehicle question in the app | Vehicle owner | Question, warning light photo, or voice prompt → Intent bound to the owner's VIN | Policy controlled |
| 2. Resolve VIN configuration and entitlements | Vehicle identity service | VIN and ownership record → BuildConfiguration and entitlement scope | Policy controlled |
| 3. Retrieve trim-valid manual and campaign content | Retrieval agent | Question plus option codes and model year → Applicable manual sections and open campaigns | Policy controlled |
| 4. Interpret telemetry and explain the condition | Vehicle companion | Latest telemetry and diagnostic codes → Plain-language explanation with urgency and cited manual text | Policy controlled |
| 5. Confirm service booking or remote command | Owner with dealer confirmation | Proposed appointment or remote action → Confirmed booking or executed command with audit record | Human required |
Technical architecture

Cosmos DB data model
| Container | Hierarchical partition key | Documents | Search paths |
|---|---|---|---|
vehicleOwnerships | /tenantId, /vehicleVin | VehicleOwnership, BuildConfiguration, TelemetrySnapshot, ServiceHistoryEntry, CampaignNotice | Operational fields only |
vehicleManualLibrary | /tenantId, /manualId | OwnerManualSection | /bodyVectorEmbedding, /bodyVector |
Query contract
| Query | Container | Scope | Purpose |
|---|---|---|---|
| Vehicle context load | vehicleOwnerships | single-partition | Read configuration, latest telemetry, and service history for the VIN in one partition query. |
| Trim-applicable manual retrieval | vehicleManualLibrary | single-partition | Knowledge-container query against vehicleManualLibrary (partition-aligned to /tenantId + /manualId): return only manual chunks valid for this VIN's option codes and model year, using the manualId resolved from the build configuration so the read is a single partition query. |
| Open safety campaigns | vehicleOwnerships | single-partition | Surface unresolved recalls for the VIN with remedy availability. |
| Service due projection | vehicleOwnerships | single-partition | Compare current odometer with the next service interval to propose a booking window. |
Python and C# services route each document type to its declared container and hierarchical partition key. ETag-based writes protect mutable aggregates, while append-only records and bounded cross-partition queries preserve scale.
Azure deployment

Clone the source repository, authenticate with Azure, and run:
Set-Location accelerators/automotive-mobility/vehicle-owner-companion
az login
./Deploy.ps1 -Runtime python -Location eastus2
Change python to csharp to deploy the .NET implementation.
Explore the implementation
- Launch signed-in demo
- Implementation article
- Domain data model
- Workflow and API contract
- Python application
- C# application
- Domain container Bicep
- Platform Bicep
- Cost and configuration
- Source repository