Skip to main content
BAN-03 · Banking & Capital Markets

Regulatory Change Impact Analyst

Outcome​

Turn incoming regulatory publications into a traceable impact assessment across internal policies, controls, and product documents, so compliance owners approve remediation work with evidence of what changed and why.

Portfolio attributeValue
Primary personaCompliance analyst
Primary aggregateRegulatoryChangeAssessment
Cosmos DB containers2
Grounding dataregulations, policies, controls, product documents
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
ProductivityImpact analysis cycle time8-15% reduction(baseline median analysis minutes - assisted median analysis minutes) / baseline median analysis minutes * 100
Quality or riskEvidence discovery time10-20% 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: Compliance analysts can compare regulations, policies, controls, and products faster without delegating regulatory conclusions.

Pilot guardrail: Version sources and escalate legal or compliance interpretation when evidence conflicts.

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 impact analysis 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. Ingest and classify the publicationRegulatory intake serviceRegulator feed document or bulletin → Classified instrument with applicability scopePolicy controlled
2. Extract atomic obligations with clause referencesExtraction agentInstrument full text → RegulatoryObligation set with deadlinesPolicy controlled
3. Match obligations to policies and controlsMapping agentObligations, policy clauses, control inventory → ControlMapping set with gap summariesPolicy controlled
4. Draft the remediation planCompliance analystGap summaries and business-line impact → RemediationItem backlog with target datesPolicy controlled
5. Accept impact assessment and planAccountable compliance ownerAssessment pack with gaps and plan → Approved plan issued to change ownersHuman required

Technical architecture​

Technical architecture

Cosmos DB data model​

ContainerHierarchical partition keyDocumentsSearch paths
regulatoryChangeAssessments/tenantId, /regulationIdRegulatoryChangeAssessment, RegulatoryObligation, ControlMapping, RemediationItem/obligationVectorEmbedding, /changeSummaryVectorEmbedding, /obligationVector, /changeSummaryVector
policyStandardsLibrary/tenantId, /policyIdInternalPolicyClause/clauseVectorEmbedding, /clauseVector

Query contract​

QueryContainerScopePurpose
Obligation to control coverageregulatoryChangeAssessmentssingle-partitionRead every obligation with its mapped controls for one instrument in a single partition.
Uncovered obligation gap listregulatoryChangeAssessmentssingle-partitionIsolate obligations whose control coverage is weak or missing.
Deadline exposure across business linesregulatoryChangeAssessmentscross-partition-boundedBounded cross-partition control-plane query: show obligations affecting a business line that fall due inside the planning horizon. It cannot carry the second partition key because a business line is affected by obligations drawn from many different regulatory instruments, so it runs as a tenant-scoped scheduled sweep off the interactive path, bounded by one business line and a 12-month compliance deadline horizon, executed with MaxItemCount paging and continuation tokens.
Stale policy clause reviewpolicyStandardsLibrarycross-partition-boundedKnowledge-container query against policyStandardsLibrary (bounded fan-out across the /tenantId + /policyId partitions named by the shortlist): flag internal policy clauses mapped to this instrument that have not been reviewed inside the governance cycle, fanned out over the policy ids the control mappings reference.

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/banking-capital-markets/regulatory-change-impact-analyst
az login
./Deploy.ps1 -Runtime python -Location eastus2

Change python to csharp to deploy the .NET implementation.

Explore the implementation​

Return to all accelerators