Skip to main content
AER-01 · Aerospace & Defense

Technical Maintenance Copilot

Outcome​

Get the maintenance engineer from a reported fault on a specific tail number to the correct effective technical order procedure for that aircraft's exact configuration, with airworthiness directives, service bulletins and parts availability resolved before the task card is released.

Portfolio attributeValue
Primary personaMaintenance engineer
Primary aggregateAircraftMaintenanceContext
Cosmos DB containers2
Grounding datatechnical orders, manuals, configurations, faults
Development estimateUSD 60-220 monthly dev estimate
Runtime optionsPython FastAPI and C# ASP.NET Core

Productivity hypothesis​

These measures reflect the workflow, evidence, decision boundary, and industry risk for this use case.

Impact typeMeasureConservative internal assumptionCustomer measurement
ProductivityTask cycle time10-20% reduction(baseline median task minutes - assisted median task minutes) / baseline median task minutes * 100
Quality or riskEvidence discovery time15-25% 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: Technical orders, manuals, configurations, and fault records can shorten maintenance preparation while preserving human maintenance authority.

Pilot guardrail: Verify configuration applicability and require maintainer approval for every consequential action.

Customer validation plan​

  1. Baseline the current workflow with representative users and tasks.
  2. Run the accelerator on the same task mix with human review retained.
  3. Measure task cycle time and evidence discovery time with the documented formulas.
  4. 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​

Domain workflow

StageActorInput → outputApproval
1. Fault intake and ATA classificationMaintenance controllerFaultReport → classifiedFaultPolicy controlled
2. Configuration effectivity resolutionConfiguration resolverConfigurationItem → effectiveProcedureSetPolicy controlled
3. Directive and bulletin compliance checkAirworthiness compliance serviceDirectiveCompliance → blockingDirectiveSetPolicy controlled
4. Task card assembly with parts and toolingTechnical maintenance copiloteffectiveProcedureSet → TaskCardDraftPolicy controlled
5. Certifying staff release and sign-offLicensed certifying engineerTaskCardDraft → releasedTaskCardHuman required

Technical architecture​

Technical architecture

Cosmos DB data model​

ContainerHierarchical partition keyDocumentsSearch paths
aircraftMaintenanceContexts/tenantId, /tailNumberAircraft, ConfigurationItem, FaultReport, DirectiveCompliance, TaskCardDraft/symptomVectorEmbedding, /symptomVector
technicalOrderLibrary/tenantId, /ataChapterTechnicalOrderChunk/chunkVectorEmbedding, /chunkVector

Query contract​

QueryContainerScopePurpose
EffectiveProcedureRetrievaltechnicalOrderLibrarysingle-partitionRetrieve only technical order content whose effectivity expression matches this tail's configuration baseline and revision level.
OpenDirectivesBlockingReturnaircraftMaintenanceContextssingle-partitionList airworthiness directives and bulletins that are due or overdue for this tail.
RecurringFaultPatternaircraftMaintenanceContextssingle-partitionDetect whether this symptom has recurred on the tail, which changes the troubleshooting path.
LifeLimitedComponentsDueaircraftMaintenanceContextssingle-partitionSurface installed components approaching life limits that could be addressed in the same hangar visit.

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​

Azure deployment

Clone the source repository, authenticate with Azure, and run:

Set-Location accelerators/aerospace-defense/technical-maintenance-copilot
az login
./Deploy.ps1 -Runtime python -Location eastus2

Change python to csharp to deploy the .NET implementation.

Explore the implementation​

Return to all accelerators