IBM CEO says AI triggers need for new operating models

International Business Machines CEO Arvind Krishna says maximizing returns on investment in artificial intelligence requires a fundamental restructuring of business workflows rather than just adopting new technology.

The use of AI within a company typically evolves from the individual contributor to small teams, cross-functional groups and ultimately the entire organization. As a company progresses from one stage to another, so do the potential returns on AI investment. That’s less about technology alone than it is about updating age-old processes and social dynamics, according to Krishna.

“In the next year or two, the enterprise world will sort into two camps: companies where AI runs their business, and companies where AI is still a project,” Krishna said.

The line between those companies that make broader use of AI and those that don’t won’t simply come down to technology. “It will be their operating model,” Krishna said.

IBM on Tuesday at its Think conference in Boston plans to announce a slew of products and capabilities, including a new version of watsonx Orchestrate, a secure multiagent control plane, and IBM Bob, for securely building and deploying agents.

While model developers are competing to stay one step ahead of each other, IBM is approaching AI from a different angle, helping its clients scale their AI efforts. AI is a core part of the technology giant’s strategy.

Last month, IBM reported higher first-quarter revenue of $15.92 billion and higher profit, driven by growing adoption of artificial-intelligence tools. While the numbers were ahead of expectations, the stock price fell. To some extent, that reflects the fact that IBM has been caught up in broader AI-driven concerns about software, but Ben Reitzes of equity research and consulting firm Melius Research has a buy rating on the company. “I’m excited…to see their AI-related business come of age,” he said.

Client zero

The use of AI at many companies began in earnest in recent years with proofs of concept. The experimentation stage has led to the greenlighting of many projects. Now companies are looking to move to the next stage, or enterprisewide deployment of AI. That’s potentially more rewarding, but also much more complex from an organizational perspective.

Krishna cited the evolution of IBM’s internal human-resource processes as an example of the operating model behind enterprise-level scaling.

In the pre-AI era, Krishna said, if an employee requested an employment verification letter to support an apartment rental application, the workflow required up to 18 different human touchpoints, including a manager, an HR business specialist, back-office staff and multiple software systems.

Today, an employee can generate the letter by making a request to an internal bot called “Ask HR.” The AI agent, integrated into IBM’s security network, automatically verifies the employee’s identity, pulls the required data from the HR system and sends the letter. All the employee has to do is specify how the letter should be delivered. The 18 touchpoints have been reduced to just one, according to Krishna.

To enable AI to scale across the enterprise, processes need to be redesigned end to end, according to Krishna. IBM’s Project Bob, as the initiative was known internally before it became a product, was designed to manage the entire software development life cycle, from writing new code to patching old code, generating documentation, creating test cases and ensuring security compliance, he said.

“I don’t begin with eliminating steps. I begin with how many touch points can I take out? And how can I make it much more nimble and faster and end-to-end? That’s the goal. Out of that comes the fact that you should eliminate steps,” Krishna said.

If people lose their jobs as a result, “I get to redeploy them, to do something else of more value,” Krishna said.

IBM has made progress rethinking its operating models, but more work lies ahead.

“I think it’s early days. We’re only a third of the way through what can be done,” he said.

The hard part

Elevance Health Chief Digital Information Officer Ratnakar Lavu said AI is transforming the way the insurance giant works. The company, a longstanding IBM customer, is working with IBM to deploy AI-driven digital assistants. IBM said it is also one of the providers that helps Elevance with AI applications such as claims and approvals. Elevance works with other AI companies on a range of applications, too. For example, it has also rolled out an internal, OpenAI-powered tool called “Spark” to help its workforce operate at peak productivity. And it applies AI to the claims and approvals process.

Through a virtual assistant, the insured can ask complex questions about their benefits, such as whether knee pain treatments are covered, and receive instant, cost-optimized provider recommendations, according to Lavu.

Like Krishna, Lavu said the successful deployment of such AI applications demands a careful recalibration of business processes across the firm. In his experience, that effort requires a deep collaboration between business and technology teams. And instead of thinking about the work as an IT project, the insurer is focused on the end-user experience. Rigorous AI governance that integrates bias testing, transparency and explainability are part of the effort from the start. As solutions are built and deployed, a parallel governance process takes place, making sure they perform ethically and within strict enterprise guidelines, Lavu said.

There are plenty of challenges along the way. Official documentation hardly ever matches the reality of how work is actually done. Businesses are bogged down by deeply embedded business rules and legacy systems. And redesigning a process with AI requires establishing entirely new checks and balances to ensure the effort doesn’t inadvertently skew key performance indicators or lead to unintended outcomes, Lavu said.

Those redesigned processes need to be connected to one another, too. For example, he said, the newly redesigned prior authorization process must continuously communicate with the newly redesigned benefits process. Only by connecting such end-to-end workflows can a company streamline operations, eliminate bottlenecks and see the full realization of AI investments, according to Lavu.

And while Elevance is seeing significant success and clear ROI in the redesign of the individual process around AI, the company “still has work to do in the connectivity of processes to see the net outcome.”

That’s the end-to-end connectivity approach driving IBM’s work. And while IBM has made progress rethinking its operating models, more work lies ahead, according to Krishna.

“I think it’s early days. We’re only a third of the way through what can be done,” he said.

Write to Steven Rosenbush at steven.rosenbush@wsj.com