900 GenAI Projects Down, J&J Says This Is How You Scale

By mid-2022, Johnson & Johnson was already deep into the generative AI boom. The company embraced an open-ended approach which CIO Jim Swanson called a “let a thousand flowers bloom” strategy giving employees across R&D, commercial, HR, and supply chain teams the freedom to propose and build GenAI use cases. A centralized governance board evaluated and greenlit nearly 900 projects, ranging from internal productivity bots to experimental drug development tools.

That phase is now officially over.

Swanson, Executive Vice President and Chief Information Officer at Johnson & Johnson, said the company has since pivoted from widespread experimentation to a focused, value-driven model. “That was a pivot we made after about a year of learning,” he said. “Now we’ve moved from the thousand flowers to a really prioritized focus on GenAI.”

The change came down to results. Internal assessments revealed that only 10% to 15% of overall AI use cases delivered any meaningful business value and those few were responsible for nearly 80% of the total impact. Many of the other projects were duplicative, difficult to scale, or better served by non-GenAI solutions.

Swanson put it bluntly: “You had to take an iterative approach to say, ‘Where are these technologies useful and where are they not?’ And I’m still a believer that there’s a lot more hype than there is substance.”

In its place, individual business units—commercial, supply chain, research, now own their own AI agendas. The logic: those closest to the work are best positioned to decide where GenAI should be applied. The company has also begun shutting down redundant pilots and consolidating resources around high-value initiatives.

Among the projects that survived this shakeout is Rep Copilot, a GenAI tool designed to coach sales reps on how to communicate with healthcare professionals. Initially piloted in J&J’s Innovative Medicine segment, it’s now expanding into MedTech, which includes devices like hip replacements and intraocular lenses.

Another investment is an internal chatbot that answers employee questions about company policies and benefits offloading some of the 10 million annual interactions currently handled by the services team. In R&D, researchers are testing whether GenAI can help identify the optimal timing for adding solvents during drug crystallization, a small but tricky step in pharmaceutical manufacturing. On the supply chain side, GenAI is being explored to forecast and mitigate risks like raw material shortages.

“This is not about chasing shiny objects,” Swanson said. “It’s about real business impact.”

Swanson stressed that none of the surviving projects were selected just because they were AI-powered. “We’re prioritizing, we’re scaling, we’re looking at the things that make the most sense,” he said. “That was part of the maturation process we went through.”

Since the release of ChatGPT in late 2022, enterprise interest in generative AI has exploded with tech giants like Microsoft and Nvidia gaining trillions in market cap. But as the hype meets the hard realities of deployment, many companies are being forced to rethink their approach. For Johnson & Johnson, that meant moving away from a culture of experimentation and toward a model that treats GenAI like any other long-term business investment—one with expected returns, metrics, and timelines.

The company now evaluates GenAI projects on three fronts: how well they’re implemented, how widely they’re adopted across the organization, and the degree to which they deliver tangible business outcomes. In this new world, fewer projects will move forward, but those that do will be expected to perform.

On LinkedIn and across tech circles, posts circulated suggesting the company’s rollback was evidence that the GenAI hype cycle had peaked or that the technology had failed to deliver. A Financial Times headline recently declared this moment in AI to be “slopaganda,” and even tech leaders like Microsoft’s Satya Nadella and IBM’s Arvind Krishna have begun voicing concerns about inflated expectations.

Instead of spreading its bets thin, Johnson & Johnson is channeling resources into fewer, better bets. Core focus areas include AI-powered drug discovery, internal productivity tools like document summarization, supply chain optimization, and sales enablement.

New areas under evaluation include clinical trial recruitment, personalized medicine through genomic analysis, and AI-assisted surgical planning. But none of these will move forward until they prove real, measurable business value.

But that’s not failure. That’s what scaling looks like in the enterprise.