How AI Is Rewriting the Corporate Playbook

Perhaps the only thing outpacing AI spending is the volume of commentary dedicated to the technology’s disruption. And while business leaders aren’t lacking for insights, they are discovering that AI is evolving faster than any coherent playbook can keep up with. The result is that companies are making high-stakes decisions without clear returns, stable organizational structures or even agreement over who’s in charge.

In a conversation with Ken Brown, The Information’s senior finance editor, Kathryn Kaminsky, U.S. chief commercial officer at PwC, described how that uncertainty is reshaping everything from investment decisions to workforce planning, and is pushing leaders to rethink not just how they deploy AI but how they define success.

Not Your Parents’ Tech Cycle

Unlike previous tech cycles where innovations primarily impacted one business function, Kaminsky observed that AI is touching every inch of the C-suite.

“The CEO obviously cares about AI because they’re investing so much. Human capital cares because they don’t know what workforce planning is going to look like. The back office cares because they’ve been the guinea pigs,” she said. “When I think about change in the world of AI, the biggest difference is how it’s impacting multiple people in the C-suite.”

From chief information officers to chief human resources officers, corporate leaders are trying to factor AI into their planning, a task the technology’s rapid evolution is making difficult. This puts pressure on traditional change management processes that weren’t designed for this pace. By way of example, Kaminsky noted the quandary CIOs are finding themselves in as they attempt to pick which models to embed in the enterprise. In the past, the decision would have been at their discretion. Today, that’s less frequently the case.

“Is the business telling the CIO what they need or is the CIO telling the business? There’s a question now of who should own the transformation,” said Kaminsky.

In practice, that uncertainty is playing out in how companies adopt AI. Rather than testing a single provider, many are testing multiple frontier models at once as capabilities evolve quickly and different tools excel at different tasks.

“CIOs normally have a process for things like what hyperscaler to use and how to use it. But what we’re seeing is that they may have to bring on more than they normally would, which in a way is very anti-ROI. How do you manage operating expenses for something like that?”

Redefining ROI

Kaminsky said that when it comes to AI, business leaders risk using an outdated lens to measure return on investment. For companies where AI poses an existential threat, delaying decisions because of ROI may be a death knell.

“The ROI of AI isn’t the issue. The ROI of doing nothing is the real issue,” she said.

With this in mind, she said, many business leaders at public companies are changing how they talk to their boards about ROI and reconsidering how they categorize those investments—blurring the line between operating expense and long-term investment.

“Taking a bet on where things are going is better than not doing anything at all,” she said. “The people who do it best are those that make assumptions as to what could happen and plan around those changes.”

A New Talent Model

AI has undeniably altered the hiring landscape, triggering companies like PwC to add new job titles—engineer, data scientist—to the accountants that traditionally made up the bulk of their workforce.

“We don’t just hire accountants anymore, because the world’s changing. We have evolved quickly in this time, and so we’re really refocusing,” she said.

More human-led skills that machines can’t easily replicate—like relatability and critical thinking—are also becoming increasingly important.

“We need people that can talk to a client with empathy and understand what they’re doing and where they’re going,” she said. The days of the “yes-man,” it seems, are also numbered, as she stresses the importance of people who can ask questions and challenge human and machine alike.

“A U.K. regulator we work with just said, ‘You can’t blame the AI.’ You still need a human lens. We really need people that will challenge us and look at the output to review it and have a dialogue around the review process,” she said. “We like to be challenged—those are the type of people we’re looking for.”