Are classical management structures holding back AI in pharma?

For decades, the pharmaceutical industry has been defined by a culture of scientific rigour and slow and cautious action. And for good reason. Long development cycles, stringent regulation, patient safety, and significant financial risk, make deliberation, control and certainty essential. 

But AI is changing the rules. When it comes to AI, lifecycles don’t run on multi-year timelines. Things change from week to week and even day to day. Models are improving continuously and capabilities evolve mid-project. That means that the organisations capturing most value aren’t waiting for perfect certainty before acting. 

This naturally creates a tension between the historically slower and more cautious cultural DNA in pharma, and a technology that rewards speed of action and learning, and adaptability.  

In this article, part of our series exploring the human side of AI transformation, we want to explore why this tension matters, and why leaders need to shift cultural priorities if they want AI to deliver more than siloed projects and short-lived pilots.

“We’re a science company, not a tech company”

As Rishi Gulati, Vice President and CIO at Otsuka, puts it in our whitepaper: “Some people get consumed with the thought that ‘we’re a science company, not a technology company’. What they are failing to understand… is that the technology is in service to the science they are going after.”

On the surface, this seems reasonable. Science is the core mission, while the tech is the support function. But given how AI is reshaping how knowledge is generated and decisions are made, treating it as a peripheral capability rather than a structural shift creates a dangerous mismatch between ambition and execution. 

And this mismatch shows up most clearly in how organisations are structured, governed and led.

Why classical management approaches may no longer be working

While traditional pharma operating and leadership models have been developed to minimise risk, they may not always work when it comes to AI. 

That’s because the technology lifecycle is moving much faster than classical management processes can accommodate. By the time initiatives pass through extended planning and governance cycles, the tools, models and possibilities have often already changed. 

And this typically leads to teams getting stuck in familiar patterns with pilots that don’t scale quickly enough, interesting prototypes that don’t embed into day-to-day work, and learning that happens too slowly. 

Most organisations have access to broadly the same technology, so that isn’t usually the issue. The differentiator is the speed of adaptation and how quickly people are able to collaborate, experiment, and adjust course as they learn. 

Why AI needs a structural transformation, not just a deployment

One of the central arguments in our whitepaper, Beyond the Algorithm, is that AI must be treated as a structural and cultural transformation. 

That’s because when it’s just treated as a deployment, organisations focus primarily on the tools, governance, and risk mitigation. These are all important things to cover, but they’re not what makes the ultimate difference when it comes to performance. By considering AI through a structural and cultural lens, leaders can really start focusing in on: 

• How work actually gets done across different functions and how they interact and collaborate

• How decisions are made and by whom

• How quickly learning happens

• How much autonomy teams have to experiment 

• Whether people feel psychologically safe enough to try, possibly fail, but learn and adapt as a result

And this is all where cultural DNA becomes decisive. Rigid, centralised models often struggle to keep pace with AI. Increasingly, organisations need teams with the permission to autonomously test, learn and iterate within clear guardrails.

As Bryn Roberts, SVP & Global Head of Data, Analytics & Research at Roche, describes: “Beyond our deeper scientific and technical use-cases, we encourage ‘everyday AI’ through playful exploration by everybody. Once people appreciate what’s possible, a little training and a safe environment enables them to ‘go and play’ in their daily work.

This doesn’t mean abandoning rigour or compliance. Quite the opposite. It means being really explicit about where precision is non-negotiable, and where experimentation is essential.

As we explore in the whitepaper, some leaders adopt a ‘care and dare’ leadership approach. Care, where mistakes carry significant cost or compliance risk. Dare, where speed of learning and innovation matter most. 

Building a culture that moves AI at speed 
AI is reshaping how knowledge is created, decisions are made, and work gets done across the pharmaceutical sector. Organisations that capture the most value aren’t those with the most advanced technology, but those whose people learn, adapt, and collaborate fastest.

Leaders play a decisive role in shaping this environment by modelling curiosity, encouraging experimentation, and creating a culture where teams are empowered to explore, learn and improve. 

Those able to embrace the shift will build a durable competitive advantage, while those who resist will quickly fall behind. The choice is clear: stay on the sidelines, or lead the transformation. For practical guidance on how to shift culture, accelerate learning, and unlock AI’s full potential, download the full whitepaper here.