Amazon Web Services on Tuesday unveiled a new set of artificial intelligence (AI) tools and capabilities designed to help companies get more value from generative AI. Amazon Web Services (AWS) said this includes a new class of AI agents that can perform tasks for hours or days without getting stuck or asking users for help.
"This is what we've been working hard on for the past year: how to build a really powerful brain that can handle complex workflows," AWS CEO Matt Garman said in an interview.
Agents are AI tools that take certain actions on behalf of users. But in their current form, they often hit bottlenecks and need to constantly look to users for guidance.
Garman said this new "leading edge agent," capable of performing a broader range of tasks for longer periods of time, is the product of what he calls "massive amounts" of software engineering and infrastructure data, a combination of different models, and a powerful underlying memory architecture.
AWS also shared news about a service called Nova Forge, which lets enterprises train private instances of Amazon's Nova models on their own proprietary data, and announced the general availability of its Trainium3 AI chip.
AWS is sometimes criticized as a laggard in AI because it has been slow to release its own models. While Garman acknowledged that at the beginning of the AI boom, AWS "took a half-step back" to focus on building a broad and scalable platform for enterprises, he said the agent trend has brought the cloud giant back to center stage. That's because agents need access to business data, core applications and other systems, most of which are supported by AWS's architecture, he said.