Inside AWS’ support for customer AI adoption: “Boots on the ground”

Enterprise generative AI adoption has been rising rapidly over the last year, with organizations globally seeking to capitalize on the potential benefits of the groundbreaking technology.

More than half of firms are using generative AI tools in some capacity now, according to research from Gartner in October. Nearly half (45%) of firms polled by the consultancy said they are using AI tools in a pilot phase, while 10% have fully implemented the technology within daily workflows.

This comes in stark contrast to previous research from the firm in March, which found that 3% of companies had begun implementing the technology. But while adoption rates continue to rise, many firms struggle to identify tangible use cases for the technology and are concerned about returns on investment. Similarly, uptake is being hampered due to the cost of implementation, as well as a lack of relevant skills or a clear-cut strategy to accelerate adoption.

To address these issues, AWS launched its Generative AI Innovation Center in July 2023. The cloud giant invested $100 million in the center, which aims to support customers during the tentative early adoption phase by connecting them with its own artificial intelligence (AI) and machine learning (ML) experts.

Embedded AI adoption teams

Industry counterparts such as Microsoft and Google both offer advice to customers in some capacity to support adoption. However, AWS made the decision to embed its own AI experts within enterprises to work in collaboration with internal teams. Teams at the center include experts from a variety of backgrounds such as enterprise strategists, data scientists, engineers, and solutions architects, all of whom work in close collaboration with customers during the experimentation process.

Sri Elaprolu

Elaprolu has worked in people management experience for more than 17 years and has experience with technical strategy going back 25 years. He has been at AWS since January 2013 and entered into his current role in May 2023.

This approach is a unique take on customer support for AI adoption, according to Sri Elaprolu, global head of the Generative AI Innovation Center. Elaprolu tells ITPro that by employing a “boots on the ground” tactic, AWS hopes to accelerate the uptake of generative AI tools among customers.

“Customers are interested in doing things with generative AI, but not a lot of them have experience or expertise in-house,” he says. “They’re wondering how they’re going to do it.”

“Many of them approached AWS for help, asking for guidance and asking how they can take advantage and how they can benefit from that technology.”

Elaprolu says that, since its launch, more than 1,000 customers have approached the firm for support. These companies span a range of industry segments across both the private and public sectors.  Much of this has been due to growing c-suite excitement around generative AI, as almost half of executives believe generative AI will boost financial performance. Part of the support offered by AWS includes connecting strategists with technology leaders to help them curate a clear-cut vision for how they can begin their AI transformation process.

“At the innovation center, there’s a strategy function,” he explains. “That team engages with those customers, understands their business direction, vision, and objectives, and then works backward from that outcome and vision to identify what they can do with generative AI to help speed up that direction.”

This process involves identifying potential use cases to establish how they can derive the most value from adoption. “We want to make sure we’re working on use cases that have the highest value and can be achieved in a practical way,” he says.

Thereafter, the innovation center connects customer teams with its own solutions architects to begin developing solutions. This typically involves leveraging the Amazon Bedrock platform to help customers tinker and experiment with specific models, such as Anthropic’s model Claude. The scaling and integration of generative AI tools is also supported during this process, Elaprolu confirms, resulting in a complete point-to-point support scheme.

“That’s what the center does - to go from ideation to prioritization, to execution, and then scaling. Along all of these steps we’re working with customer teams, both business and technical to make sure that what we’re doing is aligned with their objectives.”

Customer AI adoption successes 

Citing a specific example, Bridgewater Associates, Elaprolu says innovation center teams helped the hedge fund manager build an investment analysis solution using Anthropic’s Claude large language model.

“What the solution does is allows you to bring data to the tool, let it decipher and identify dependencies to visualize that data,” Elaprolu explains. “For example, if you’re looking at a macro investment portfolio and inflation in any part of the globe changes – up or down – it analyzes and visualizes that.”

The PGA Tour is another major customer the innovation center has worked with so far. Together with AWS, the sports organizer is creating an intuitive chatbot capable of natural language processing (NLP) aimed at improving user experiences on its website. The development process here brought together a range of disparate datasets so that PGA Tour could produce better insights according to a range of variables.

“PGA Tour sits on a lot of data,” he says. “They have data about players, about performance, about matches and events. There’s a wide range of data sources that PGA Tour collects and maintains and wants to make available for users in an easy-to-interact way as opposed to going and searching.”

“We helped them build a chat interface that allows you to get lots of interesting details out of these disparate data sources in plain English.”

This has received “a lot of positive feedback” from both PGA itself and customers, Elaprolu adds.  AWS’ current approach to AI adoption support follows in the example of its work providing support for customers embarking on cloud migration and transformation projects, or overhauling IT architectures.  Elaprolu says it’s something that AWS has been “doing from the beginning” across other segments of the business, and as such made sense to apply to soaring demand for generative AI.

Read more from AWS re:Invent

• AWS re:Invent 2023: All the day-three keynote announcements
• Amazon CodeWhisperer updates could be a game changer for shifting left
• New Amazon S3 Express One Zone promises 10x performance boost
• AWS unveils ‘Amazon Q’ enterprise-grade AI assistant
• Amazon Q's foundations separate it from ‘monolithic’ competitor AI tools
• Amazon Detective offers security analysts their own generative AI sidekick