AWS taps agentic AI to slash cloud migration time from months to days

  • AWS is developing a new migration tool that can tackle some of the hardest parts of mainframe and VMware migrations
  • Amazon Q Developer can automatically update code and help manage project tasks
  • There are plenty of competitors in the market, analyst Roy Chua noted

Forget headaches. Migrating ancient mainframes and VMware workloads to the cloud can be outright migraine-inducing projects. But Amazon is swooping in with a much-needed painkiller in the form of its new AI-powered Amazon Q Developer tool.

Though the tech is still in preview, it is already showing wild promise. Some beta customers have been able to slash migration timelines from as much as eight months to just a few days, AWS VP of Technology Mai-Lan Tomsen Bukovec told Fierce.

According to Bukovec, Amazon Q Developer comprises a unique technology stack that incorporates AI agents and an agentic AI orchestrator. The AI has been trained to deeply understand mainframes, VMware networks and AWS environments. But, Bukovac noted, “it’s not all AI.” There’s also a graph neural net for network mapping, algorithmic code and additional capabilities sourced from foundation models.

The way it works is simple: Basically, Amazon Q Developer takes over all the rote migration tasks like code updates (for instance, from COBOL to Java or .NET to .NETcore) and project management tasks (like generating to-do lists, tracking approvals and completion status), but keeps a human in the loop to sign off on its work. And the beauty of AI means it can run hundreds workload projects in parallel to save even more time.

This frees up “highly constrained talent” and means IT departments can more readily kick off migrations without needing to ask app developers in different departments across the enterprise for their time, she stated.

Put it all together and you’ve got one heck of a migration machine. Bukovec said Amazon Q Developer represents the “next gen of collaboration” between humans and AI to tackle complex projects.

She added that as part of the preview process, AWS is working with the likes of Accenture, Amdocs and McKinsey – who are commonly called in to help with migrations – to generate feedback on how to improve Amazon Q Developer before it heads to general availability.

“We’re spending a lot of time right now getting feedback from all of those different team members to make sure we’re providing the right set of capabilities,” she said. “That whole collaboration surface we think is really important to get right.”

As you might expect, Bukovec said AWS doesn’t really believe it has any competition out there with the capabilities it offers. But AvidThink founder Roy Chua countered that migrations are big business for several others.

For instance, he noted many of the companies AWS is working with on the Amazon Q Developer preview “also have in-house proprietary tools and internal best practices — the system integrators are both partners and competitors in a sense.” Case in point: Amdocs, which bought Astadia in 2023 specifically to address migration hurdles.

Then there’s IBM, Chua said, which has been demonstrating how to use their watsonX AI tools to do the same. Ditto companies like Kyndryl, Ensono and Accenture.

That said, Chua said creating a tool like Amazon Q Developer is a “smart move to drive more workloads to the cloud, a good fit for generative AI's code gen[eration] capabilities and reduces the hesitancy on the part of companies to transition to the cloud.”

Fierce will be watching Amazon Q Developer closely. Why? Because of the implications it could have for telcos specifically. Bukovec said AWS has seen a lot of interest in using the tool to move VMware applications specifically. And as we noted previously, telecom is a rather large VMware estate.

Our spidey senses are tingling.