Fleet Operations Copilot
Outcome
Keep a mixed combustion and electric fleet running by fusing telemetry, route plans, charging availability, and maintenance status into ranked interventions, with the fleet manager approving vehicle swaps and route changes.
| Portfolio attribute | Value |
|---|---|
| Primary persona | Fleet manager |
| Primary aggregate | FleetOperationsDay |
| Cosmos DB containers | 1 |
| Grounding data | telemetry, routes, charging, maintenance, weather |
| 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 | Operational triage time | 10-20% reduction | (baseline median triage minutes - assisted median triage minutes) / baseline median triage minutes * 100 |
| Quality or risk | Rework rate | 5-10% reduction | (baseline reopened, corrected, or repeated task rate - assisted rate) / baseline rate * 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: Fleet managers can reconcile telemetry, routes, charging, maintenance, and weather faster for bounded operational decisions.
Pilot guardrail: Do not override dispatch, safety, routing, maintenance, or charging policies.
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 operational triage time and rework rate 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. Build the depot day plan | Fleet planner | Route commitments and vehicle roster → FleetOperationsDay with assignments and energy plan | Policy controlled |
| 2. Ingest vehicle telemetry and charger state | Telematics integration service | State of charge, faults, charger availability → Refreshed VehicleStatus and charging capacity | Policy controlled |
| 3. Detect range, maintenance, and capacity exceptions | Exception detection agent | Route energy demand versus vehicle and charger supply → OperationalException set with severity | Policy controlled |
| 4. Recommend swaps, charging, or re-routes | Fleet copilot | Exceptions, weather, and maintenance constraints → Ranked remedies with cost and service impact | Policy controlled |
| 5. Approve the dispatch change | Fleet manager | Proposed vehicle swap or route change → Committed dispatch update sent to drivers | Human required |
Technical architecture

Cosmos DB data model
| Container | Hierarchical partition key | Documents | Search paths |
|---|---|---|---|
fleetOperationsDays | /tenantId, /fleetDepotId | FleetOperationsDay, VehicleStatus, RoutePlan, ChargingSessionPlan, MaintenanceDueItem, OperationalException | /remedyVectorEmbedding, /remedyVector |
Query contract
| Query | Container | Scope | Purpose |
|---|---|---|---|
| Depot readiness view | fleetOperationsDays | single-partition | Read vehicle status, routes, and charging plans for the depot day in one partition query. |
| Range risk detection | fleetOperationsDays | single-partition | Find electric vehicles whose charge cannot cover the assigned route energy demand. |
| Dispatch-blocking maintenance | fleetOperationsDays | single-partition | List vehicles that cannot legally or safely be dispatched today. |
| Unapproved exception queue | fleetOperationsDays | single-partition | Show remedies awaiting the fleet manager's decision before departure. |
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/fleet-operations-copilot
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