From Manual Drafting to AI-Powered Agility: FIDUMTEC’s Success Story

FIDUMTEC is a privately-held IT and telecommunications firm headquartered in Buenos Aires, Argentina, with operations across Latin America and the U.S. Founded in 2016, the company specializes in big data, IoT, machine learning, real-time communications, and cybersecurity. With a multidisciplinary team and presence in more than eight countries, FIDUMTEC serves both public and private sector clients, delivering scalable, high-impact systems tailored to regional needs. Its agile delivery model and domain expertise position the company as a strategic partner for innovative solutions powered by generative AI.

The Business Challenge

FIDUMTEC faced growing inefficiencies in the early stages of its agile delivery lifecycle. The quality and consistency of user stories and test cases varied significantly across teams, leading to ambiguity, misalignment, and costly rework. Drafting these artifacts manually was slow, heavily dependent on the analyst’s expertise, and often left gaps in coverage. As FIDUMTEC scaled agile projects across multiple industries, it needed a solution to standardize, accelerate, and enhance requirements documentation—without sacrificing clarity, functional depth, or testability.

The Solution

To solve FIDUMTEC’s challenge, DIRMOD implemented a Generative AI assistant powered by Amazon Bedrock. The solution integrates seamlessly with the client’s agile workflows and leverages Anthropic and Amazon Titan foundation models via a custom prompt-engineering layer. Using retrieval-augmented generation (RAG), the assistant contextualizes project-specific documentation—such as epics, business rules, and historical test cases—to automatically generate consistent, high-quality user stories and test cases.

The architecture was deployed serverlessly on AWS Lambda, Amazon API Gateway, and Amazon S3, ensuring scalability, security, and cost efficiency. The assistant delivers editable, traceable outputs that support collaboration across analysts, developers, and QA engineers—accelerating early development cycles and strengthening alignment with agile standards.

Results and Benefits Achieved:

The collaboration between FIDUMTEC and DirMOD resulted in:

  • 40–60% reduction in time to draft user stories and test cases.
  • 30% improvement in coverage and consistency of requirements, validated by internal QA audits.
  • 25% increase in early defect detection due to clearer acceptance criteria.
  • 90% user satisfaction among business analysts and QA engineers, citing time savings, ease of use, and alignment with agile best practices.
  • The solution is now being expanded across delivery teams, with future integration planned for sprint planning and regression testing workflows.

“With DIRMOD and AWS, we transformed one of the most time-consuming parts of our agile process. Generative AI now helps our teams deliver clearer, more consistent user stories and test cases in a fraction of the time—boosting both quality and speed. This innovation allows our analysts and engineers to focus on what truly adds value for our clients.”

— Hernán Galdano, CEO, FIDUMTEC