Utilizing foundation models to drive enterprise ROI

A successful AI strategy has become critical to drive efficiency, avoid stagnation, and stay ahead of competitors. Amazon‘s Vaughn Schermerhorn, IBM‘s Rob Thomas, and NVIDIA‘s Shanker Trivedi — moderated by NYSE’s Chuck Adkins — share their insights on how to integrate AI across the enterprise, where to start for the fastest ROI, and how to measure meaningful results.

These insights came from our ScaleUp: AI event in October 2023, an industry-leading global conference that features topics across technologies and industries. Watch the full session below:

Augment and enhance work — don’t just replace it

The panelists agreed that artificial intelligence is not just about replacing human jobs, but about augmenting and enhancing existing work. Trivedi described AI as a tool that changes how work is done, while Adkins emphasized the importance of using AI to bring efficiency and boost productivity.

“You need to inject humans into this process as well… without constant human involvement, these models tend to collapse.”

Trivedi highlighted the importance of domain specificity in AI models, stating, “We think that…it’s basically going to trend towards domain-specific language models and generative image and video.” Schermerhorn also emphasized the importance of human involvement in training domain-specific models. He stated, “You need to inject humans into this process as well… without constant human involvement, these models tend to collapse.”

“Imagination and willingness” to adopt AI

Panelists discussed the importance of return on investment (ROI) and prioritization in deciding which projects should incorporate AI. Thomas stressed the importance of prioritizing projects that could bring ROI to the company quickly.

“I think the greatest value we provide to other software companies is actually distribution… we’re driving go-to-market for them.”

He stated, “I think the greatest value we provide to other software companies is actually distribution… we’re driving go-to-market for them.” He also mentioned the need for companies to have the “imagination and willingness” to adopt AI.

Schermerhorn shared how Amazon uses AI to solve customer problems and justify cost expenditure. He mentioned, “Leveraged correctly, [AI] solves customer problems at scale efficiently in a way that we’re really seeing great returns for.”

Dealing with sensitive data

Thomas voiced the challenge of dealing with personally identifiable information (PII) and sensitive data. “Identifiable information, sensitive data — this is what slows down a lot of companies that are trying to figure out some generative AI. They’re basically trying to mix, you know, what I’d call data that they’re comfortable using and data they think is risky.”

Thomas also explained, “Most companies aren’t ready to figure out how do they have data in a usable format to do this. I’d say that one sticks out.”