Overview
IBM Consulting's Digital Business Analysis (BA) is an agentic AI business analysis assistant purpose-built for Communications Service Providers (CSPs). It ingests requirement specifications, RFPs, and solution design documents, then runs a multi-agent pipeline on Amazon Bedrock AgentCore to extract business rules, organize them into capabilities and features, and generate fully structured user stories and acceptance criteria, turning days of manual requirements analysis into a review-ready backlog in minutes. How It Works • Retrieval-augmented extraction: Indexes the source document using Amazon Bedrock Titan embeddings and a FAISS vector store, then extracts business rules, scope, and out-of-scope items with page-level source citations for full traceability. • Capability modeling: Classifies extracted features into TM Forum business domains — Customer Management, Service Management, Resource Management (OSS), Business Assurance & Revenue Management, and others, so output maps directly onto how CSPs structure their OSS/BSS programs. • Automated user story generation: Writes “As a / I want / So that” user stories for every in-scope feature, each linked back to its source business rules. • Built-in acceptance criteria quality gate: Enforces that every user story carries at least one acceptance criterion, and that the total acceptance criteria count exceeds the total story count preventing thin, untestable requirements from reaching the backlog. • Section-aware processing for large documents: Automatically partitions long specifications by section, runs extraction per section, then merges and deduplicates results across the full document. • One-click requirement compiler: Consolidates business rules, capabilities, features, user stories, and acceptance criteria into a single formatted Excel workbook, organized as Capability → Feature → User Story → Acceptance Criteria. AWS Components • Amazon Bedrock AgentCore Runtime: Hosts and orchestrates the multi-agent pipeline — managing agent lifecycle, routing tasks between the Orchestrator and specialist agents, and streaming results back to the UI in real time. • Amazon Bedrock AgentCore Knowledge Base: Stores indexed document chunks, TM Forum domain mappings, and extracted business context, enabling agents to retrieve relevant passages via semantic search during extraction and classification. • Amazon Bedrock Foundation Models (Claude Sonnet): Powers natural language understanding for requirement parsing, business rule extraction, user story generation, and acceptance criteria authoring across all agent steps. • Amazon Bedrock Titan Embeddings: Generates vector embeddings of the ingested document, enabling FAISS-based semantic retrieval so agents can locate relevant sections and source-cite business rules back to their originating pages. • AWS Cognito & AWS IAM: Enforce authentication and fine-grained access control for the review UI and agent APIs, ensuring all interactions are identity-verified and scoped to the customer's own AWS account boundary. Why Telecom Operators Choose Digital BA • Speaks the language of CSPs: Industry aligned output integrates naturally with existing OSS/BSS transformation, 5G rollout, and order-to-cash programs. • Deploys inside AWS account: Runs entirely on Amazon Bedrock AgentCore within your own AWS account, keeping specification documents and extracted requirements inside your security boundary. • Audit-ready traceability: Every business rule and acceptance criterion links back to its source page and originating rule, supporting compliance review and sign-off.
Highlights
- Business Value Highlights • Reducae backlog in minutes: Turns RFPs, specs, design docs into a backlog that's ready for review (business rules, capabilities, user stories & acceptance criteria), cutting BA effort from days to minutes. • Zero untestable stories: A built-in quality gate requires acceptance criteria on every story • Industry-aligned: Sorts features into TM Forum domains (Customer, Service, Resource & Revenue Management) to match how CSPs run OSS/BSS programs.
- • E2E automation: One multi-agent pipeline on Amazon Bedrock AgentCore covers ingestion, RAG rule extraction, TM Forum classification and story generation. • Audit-ready: Every rule, story and criterion cites its source page, which makes compliance and traceability reviews easier. • Built for large telecom specs: Splits long RFPs by section, extracts each section, then merges the results, removes duplicates. • Stays in your AWS account: No spec content or outputs leave your security perimeter.
Details
Pricing
Custom pricing options
How can we make this page better?
Legal
Content disclaimer
Support
Vendor support
• Saon Sen Nandi (saonandi@in.ibm.com ) • Neal Bhattacharya (nilanjan.bhattacharya@us.ibm.com ) • Debajyoti Mukherjee (debajmuk@in.ibm.com )