As the organisation's strategic Quality Enablement partner, Fabric spearheaded the design and implementation of an AI-powered test case generation platform, transforming user stories and acceptance criteria into structured, TMF-aware test scenarios.

Using Vertex AI (Gemini) and the Vercel AI SDK, the solution intelligently interpreted requirements, enriched them with TM Forum standards, and automatically generated positive, negative and edge-case test scenarios. An agentic workflow was introduced to resolve incomplete or ambiguous requirements, reducing the need for manual clarification sessions and improving the quality of generated outputs.The agent first decomposed each requirement into discrete testable claims and flagged missing preconditions, undefined actors, conflicting business rules, or terms not mapped to any TM Forum standard entity. The agent queried the TM Forum knowledge base (SID/eTOM/Open API artifacts) as a retrieval tool to infer the most probable intent — e.g. status transitions, or field constraints from the standard's canonical model.

The platform integrated directly with the existing QE workflow through Xray, enabling generated test cases to move seamlessly into review and execution while maintaining traceability between requirements and testing activities.

‍

The solution modernised the organisation’s QE process, significantly reducing manual effort while improving testing consistency and coverage across TMF-compliant APIs.

• Accelerated time-to-market by reducing the QE preparation and review cycle from 1.5 weeks to 3 days, dramatically accelerating testing readiness.
• Lowered the cost of quality by automating repetitive test design activities and reducing manual effort.
• Increased Quality Engineer productivity, enabling teams to focus on higher-value exploratory, integration and risk-based testing rather than documentation.
• Reduced rework by identifying requirement gaps and ambiguities earlier in the delivery lifecycle.
• Improved release confidence through more comprehensive and consistent API test coverage.
• Established a reusable AI capability that can be scaled across additional products and programs, delivering ongoing productivity gains without proportional increases in QE effort.
• Enabled the organisation to support increasing delivery volumes without proportionally growing the QE team.

By combining generative AI with deep telecommunications domain knowledge and TM Forum standards, Fabric helped a large telecommunications organisation establish a faster, more efficient and more intelligent approach to quality engineering—reducing operational overhead while improving software quality at scale.

‍