Skip to main content
TEL-01 · Telecommunications

Network Operations Copilot

Outcome​

Cut mean time to repair in the network operations centre by correlating alarms with topology and change records, retrieving the matching runbook, and requiring engineer approval before any automated remediation touches the network.

Portfolio attributeValue
Primary personaNOC engineer
Primary aggregateNetworkIncident
Cosmos DB containers2
Grounding dataalarms, topology, logs, tickets, runbooks
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
ProductivityAlarm-to-approved-action time10-18% reduction(baseline median alarm-to-action minutes - assisted median minutes) / baseline median minutes * 100
Quality or riskRework rate5-10% 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: Network operators can prepare diagnoses and response options faster within existing change controls.

Pilot guardrail: Do not execute network changes autonomously; require operator verification of alarms, topology, logs, tickets, and runbooks.

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 alarm-to-approved-action time and rework rate 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. Ingest and deduplicate alarmsFault management integrationAlarm stream from network elements → Deduplicated alarm set per regionPolicy controlled
2. Correlate alarms into a service incidentCorrelation agentAlarms plus topology dependencies → NetworkIncident with service impactPolicy controlled
3. Test change and maintenance causationCausation agentRecent change records in the blast radius → Suspected cause with supporting change evidencePolicy controlled
4. Retrieve and adapt the runbookRunbook retrieval agentFault signature and element type → Ordered diagnostic and remediation stepsPolicy controlled
5. Approve and execute remediationNOC engineerRemediationProposal with predicted impact → Executed change with rollback plan and incident updateHuman required

Technical architecture​

Technical architecture

Cosmos DB data model​

ContainerHierarchical partition keyDocumentsSearch paths
networkIncidents/tenantId, /networkRegionIdNetworkIncident, NetworkAlarm, TopologyRelation, ChangeRecord, RemediationProposal/impactSummaryVectorEmbedding, /impactSummaryVector
networkRunbookLibrary/tenantId, /faultSignatureCodeRunbookProcedure/stepVectorEmbedding, /stepVector

Query contract​

QueryContainerScopePurpose
Incident correlation contextnetworkIncidentssingle-partitionRead alarms and topology for the region inside the correlation window in one partition query.
Recent change blast radiusnetworkIncidentssingle-partitionFind changes touching the impacted elements shortly before the fault.
Runbook signature matchnetworkRunbookLibrarysingle-partitionKnowledge-container query against networkRunbookLibrary (partition-aligned to /tenantId + /faultSignatureCode): retrieve the runbook steps whose fault signature matches the correlated incident; the library stores one runbook document per fault signature so the lookup is a single partition read.
Priority incident boardnetworkIncidentssingle-partitionShow unresolved incidents ordered by customer-impacting priority.

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/telecommunications/network-operations-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