As the world's largest cloud provider, Amazon Web Services stands to gain the most from generative artificial intelligence.
AWS intends to do that with Bedrock, a new service that gives customers access to a variety of foundational AI models to build and customize apps using their own data. Still, most AWS clients don't quite know where to start with generative AI — that's where the AWS Generative AI Innovation Center comes in, Sri Elaprolu, the global head of AWS Generative AI Innovation Center and an 11-year AWS veteran, said.
The Generative AI Innovation Center, which Elaprolu leads, is a global team of AI and machine-learning science and strategy experts at the company. The team helps AWS customers of all sizes develop and create generative-AI applications.
Interest in the Generative AI Innovation Center has been high since the company announced it in June, Elaprolu said. The team works with over 1,000 customers, including the PGA Tour and Bridgewater Associates, to move generative-AI apps from the proof-of-concept stage to production.
"The transformation capability that generative AI is bringing is immense," Elaprolu told Business Insider. "More and more companies clearly see the value in optimizing their operations, saving costs, and bringing more innovative solutions to customers."
Most companies are starting to experiment with generative AI through applications designed to help speed up workflows they can implement internally, Elaprolu said. With the help of Elaprolu's team, Bridgewater is testing a generative-AI tool to speed up causal analysis, a type of modeling financial analysts usually must do manually. The PGA Tour created a chatbot that answers golf-history questions and gives a text or video response generated from its content archives.
Elaprolu's team works with AWS customers in all stages of generative-AI development. His only requirement is that customers come to him with a specific business outcome in mind.
"We don't want to start with a technology solution and then try to find a business problem, versus working backwards to a technical solution that addresses the problem," Elaprolu said.
Otherwise, he said, developing a generative-AI app can become an endless science experiment.
"You do these experiments with no end goal in mind, and you just waste cycles and lose momentum," he said.
Many AWS customers are just beginning to think about generative AI, but an increasing number are starting to bring ideas into production.
"It's no longer just talk and no longer just experiment," Elaprolu said. "It's now getting serious."