IBM CEO Rometty's 5 Lessons For Scaling AI

Lessons Learned

IBM has been at the forefront of bringing artificial intelligence to the enterprise through its Watson Cognitive Computing Platform.

Through that work, the tech stalwart has distilled five lessons that can help guide digital transformation initiatives, CEO Ginny Rometty told attendees at IBM's THINK 2019 conference Tuesday in San Francisco.

In her keynote at THINK, Rometty said the industry is entering Chapter 2 of "digital reinvention"—an enterprise-driven chapter that will see AI embedded everywhere in the business.

She shared IBM's five lessons to help customers succeed with those initiatives, while also reminding them that as they look to AI to drive efficiencies and engage with customers in entirely new ways, "underpinning it is trust, and that's going to be about responsible stewardship."

Outside In

The first and second lessons IBM CEO Ginny Rometty shared regard the approach enterprises take to adopting artificial intelligence.

IBM sees many of its largest customers start their AI transformations by "changing everything around the customer," Rometty said.

That typically means developing and deploying new customer experience apps and entirely new services to engage those customers.

Rometty calls that approach "pushing from the outside in."

Inside Out

Another approach to digital transformation and AI enablement starts internally with modernizing the enterprise workflow.

That's a preferred route for many IBM customers already very familiar with the idea of a workflow, and how data drives change in their internal processes, said IBM's CEO.

"For this you got to modernize your core applications," Rometty said. "But you end up with an architecture that's pretty flexible."

Business Platform

The third lesson of digital transformation is the need for a capable business platform to connect outside-in and inside-out approaches to AI adoption.

That platform will "be fueled by data," Rometty said, and "it will have AI infused in all of its different workflow."

AI Platform

The process of scaling AI throughout the enterprise becomes a virtuous circle, Rometty said, once enterprises see the need for an AI platform.

Such a platform is necessary because AI, like legacy applications, needs to keep track of its lifecycle.

"Just like old days when you needed a lifecycle development platform for applications," IBM's CEO told THINK attendees, "you're going to need an AI platform."

Information Architecture

"You will never have AI without IA," Rometty told IBM customers.

An information architecture is essential to scaling deployment of artificial intelligence capabilities in an era where 80 percent of the effort in leveraging AI involves getting data ready to use.