Evertec
AI agents across the software development pipeline
Evertec's screen-building agent cut a five-day task to 10-15 minutes
Problem
Evertec's Brazilian unit maintains large legacy systems for banks and must adapt them to regulatory changes, such as the move to alphanumeric CNPJ company registration numbers due in 2027. Screen building, documentation and code changes took many developer days.
Approach
Evertec built its own AI agents for each phase of the pipeline: specification, development, testing and deployment. For legacy systems it used four agents: screen generation to company visual standards, regulatory adaptation, code documentation, and a change agent that finds problems across screens, backend and database and applies fixes after a developer approves. The work is part of an R$30 million AI research and development investment.
Results
Takeaway
Evertec split the pipeline by phase and gave each phase its own agent, with developer approval before fixes were applied. The figures come from the company's own AI director.
Categories
- Industry
- Technology and telecom
- Business function
- IT and R&D
- Innovation type
- Process innovation
- AI technology
- AI agentsGenerative AI
- Adoption
- Built in-house
- Evidence
- Company's own claim
Sources
Source summaries
Mobile Time interviewed Evertec's AI director about the company's R$30 million AI investment and its development agents. The article gives a 60% efficiency gain in development, a screen task cut from about five days to 10-15 minutes, and legal changes made 85-90% faster. It also says 51% of professionals use AI in development and more than 400 use coding assistants daily. All figures come from the company and were not verified independently.
Full details are in the original sources.