Skip to main content
AI & Technology

Itaú and GFT cut mainframe batch migration time by up to 70%

Itaú Unibanco and GFT say AI agents cut migration time for mainframe batch systems by up to 70% and sped up delivery by up to 3.3 times in the initial phase.

  • Nov 6, 2026
  • 2 min read
  • LabForty AI Newsroom
Itaú and GFT cut mainframe batch migration time by up to 70%

Itaú Unibanco and GFT Technologies say AI agents cut the time needed to migrate the bank's mainframe batch systems to the cloud by up to 70% in the initial phase of the work, and made project delivery up to 3.3 times faster, according to a GFT press release dated October 29, 2026.

Batch systems are the scheduled jobs that process large volumes of transactions at once, such as end-of-day closings, settlements and reconciliations. At many banks they still run on mainframes, in code written decades ago. Moving them is risky because the business rules live inside that code.

The work is built on Batch2Cloud, a tool GFT developed together with the bank and integrated into Wynxx, GFT's generative AI platform for legacy modernization. It is used to find modernization candidates, map the complexity of the environment and estimate effort. Itaú also used Code Understanding, a tool from Zup Innovation, the group's technology company, to help agents interpret legacy code.

The split of responsibilities is explicit. Agents take on repetitive tasks, speed up analysis and support code generation and validation. Engineers stay responsible for technical decisions, solution architecture and governance. Douglas Santos, Technology Superintendent at Itaú Unibanco, said the agents operate under the bank's responsible use policies. GFT CEO Marco Santos said the work changes how large institutions approach mainframe modernization.

The figures are the companies' own and describe the initial implementation phase. The release does not say how many batch jobs were migrated, what baseline the 70% is measured against, or how the bank verified that migrated jobs produce identical results. It places the project inside a cloud migration plan Itaú started in 2018.

Why this matters for banks

Most bank AI announcements describe customer-facing assistants. This one targets the back end, where legacy code limits how fast new products can ship. The release describes half of a workable model: agents produce analysis and drafts, people own the architecture. The other half, proof that every migrated job produces the same results as before, is the part any bank should ask to see in detail.

This article is not legal advice.

Where every detail matters

Where every detail matters

At LabForty, we develop high-quality websites with a strong focus on detail - from architecture and user experience to business logic.