WTF Summit Recap: AI Adoption 2.0: What Comes After the Pilot Projects

After years of hype and hundreds of billions of dollars in investment, many executives are asking the same question: When will AI start paying off? Reports like MIT’s much-debated survey estimating that 95% of generative AI pilots fail suggest widespread frustration, but some companies are finally moving beyond experiments to projects with impact.

To explore what successful adoption looks like in practice, The Information’s founder and editor in chief, Jessica Lessin, sat down with Jenny Koehler, U.S. COO of Advisory at PwC. Koehler oversees one of the world’s largest consulting businesses and has a front-row view of how major corporations are using—and struggling to use—AI to transform.

AI Adoption’s Top-Down Move

In contrast to MIT’s study, Koehler pointed to PwC’s most recent CEO survey as evidence that corporations are starting to see real gains from AI investments: 56% of CEOs said they’re already realizing efficiency improvements, and one-third reported higher revenue and profitability.

“They’re starting to see those green shoots emerge,” she said.

That momentum, she added, has changed how companies approach adoption. Over the past six to nine months, many have shifted from grassroots, bottom-up pilots—where employees across the organization propose use cases—to top-down, value-driven programs aimed at improving efficiency and profit at scale.

“For our corporate clients, the projects that get the most traction and the most legs are those that hit the top line,” she said.

A Tale of Two Corporate AI Strategies

For the most part, Koehler said, her clients’ adoption strategies fall in one of two camps.

“There are the clients that are looking for efficiency—to help the operations run better—and there are those that want to get new products to market faster,” she said.

Efficiency plays vary by industry: predictive maintenance in manufacturing, automated contract review in legal, portfolio-risk analysis in consulting. In pharmaceuticals, AI’s biggest impact comes from accelerating drug discovery and clinical trials.

“The speed to insight is faster than it’s ever been,” she said.

Human Upskilling in an AI World

For adoption to succeed, Koehler emphasized, employees need both access to AI tools and the freedom to experiment with them.

“We’re trying to meet this market moment as quickly as we can,” said Koehler. “The first thing we do is democratize the technology and get it into the hands of our practitioners, who I think left to their own devices will do amazing things.”

But democratization also means disruption. The degree to which AI augments a workforce, she noted, depends largely on whether a business is growing or contracting.

At PwC, training focuses less on technical upskilling than on human upskilling.

“We are spending a lot of energy and money on the human-plus-AI dimension in our training curriculum,” Koehler said. “We’ve codified what we believe to be 15 distinctly human skills.” These include empathy, judgment, ethics, and storytelling—skills she believes will define the next generation of leaders.

Leadership in the Age of AI

The age of AI, Koehler argued, will demand a new kind of leadership—one rooted in transparency and emotional intelligence rather than authority or performance.

“The notion of a leader that’s playing a part won’t go very far in this world, which is moving so fast and where there’s so much uncertainty,” she said. “A good leader will be one who shows vulnerability, who’s honest and transparent, and [who] has the courage to hold space and say, ‘I don’t know.’”