Why Centralizing AI Is the Wrong Move for Medical Affairs

Over the past few months, I’ve had a some conversations with pharma leaders trying to define the right structure for Medical Affairs when it comes to AI, digital innovation, and data insights.

An interesting scenario is emerging: Medical Affairs is being asked to give up headcount—roles previously dedicated to Medical-specific AI, technology, and data efforts—and shift them into centralized enterprise functions that support both Medical and Commercial. Or, if they aren’t asked to give up headcount, they are promised support if they assign their AI execution to a centralized team--and can the redeploy there headcount previously assigned to Medical innovation-type roles.

And while there’s often a promise of a “dedicated Medical lead” within the centralized team, it’s usually a politically convenient compromise with strategic consequences. That lead ends up navigating two very different sets of priorities—enterprise KPIs on one side, scientific and Medical needs on the other. Without clear ownership, budget, or decision rights, Medical’s agenda inevitably gets deprioritized.

And while it may meet a political agenda for a Medical leader—appeasing the enterprise, checking the innovation box, or signaling cross-functional alignment—it sacrifices the very thing that could pull the function out of strategic obscurity. Especially now, as the industry begins to recognize that the buzz around Medical impact measurement has fizzled, the next lever for relevance isn’t just metrics—it’s capability. And that capability won’t be built from the outside in.

MEDICAL AFFAIRS IS UNDERPOWERED--BUT THAT'S A DESIGN PROBLEM, NOT A CAPABILITY PROBLEM

Let’s be honest: Medical Affairs hasn’t led the digital charge in most organizations. But that’s not because the function lacks good ideas or smart people. It’s because there hasn’t been real commitment to building the necessary infrastructure inside Medical itself.

In many cases, roles tied to innovation are handed out informally—often to whoever shows interest. A Medical Director finishes an AI course and suddenly becomes the “AI lead.” An MSL gets certified in omnichannel engagement and is now driving digital strategy. Meanwhile, these roles often have no budget, no team, and no clear mandate.

Even when Medical leaders believe in innovation, they rarely allocate serious resources to it. If asked to choose between hiring another Medical Director or investing in AI and digital innovation, most opt for the former. Why? Because innovation is hard to define, requires patience, and often involves failure. It’s safer to invest in what’s known and quantifiable.

But this pattern explains why so many Medical-led digital efforts—especially in HCP engagement—have underdelivered. The intent was there. The capacity wasn’t.

Trying to run a transformation program on a pilot-sized budget is like trying to power a city on a backup generator—it might keep the lights on, but it won’t scale. Medical innovation needs infrastructure, not improvisation.

 WHEN MEDICAL BECOMES THE CLIENT, INNOVATION BECOMES REACTIVE

Once digital, AI, and data teams are moved into a centralized enterprise function, Medical Affairs no longer owns its roadmap—it becomes a client of someone else’s. And that someone else is often incentivized by Commercial outcomes, not scientific or Medical value.

What usually follows is a dual-reporting structure. The AI lead reports to both the enterprise team (which controls tools and budget) and to Medical (which controls the needs). And that’s where it breaks down.

Progress slows. Priorities get diluted. Medical’s voice weakens. Enterprise teams may promise “equal support,” but we all know where the money goes: to the parts of the business with short-term commercial ROI.

Without a dedicated, embedded AI and digital lead fully housed in Medical Affairs, the function is left performing innovation theater—presentations instead of platforms. Pilots instead of progress.

Medical needs proactive innovation with AI; not reactive innovation.

WE'VE SEEN THIS BEFORE--AND WE KNOW HOW IT CAN END

Look no further than Target Canada. The company made centralized decisions about supply chain and inventory from its U.S. headquarters—without real insight into local operations. The result: chronic stockouts, empty shelves, and a $2 billion failure. Target exited the market entirely within two years. [1]

The same dynamics apply here. Centralizing innovation without domain knowledge and contextual sensitivity doesn’t lead to scale. It leads to collapse.

For Medical Affairs, the consequences are clear:

• Compliance friction, when tools designed for Commercial are retrofitted for Medical use

• Scientific misalignment, when models optimize for marketing KPIs instead of clinical relevance

• Operational drag, when every decision has to be escalated outside the function for alignment

THE RESEARCH BACKS IT UP: EMBED, DON'T EXTRACT

This isn’t just opinion. The literature supports it.

• Harvard Business Review warns that centralized innovation teams often “operate in silos, disconnected from the business units they’re meant to support.” [2]

• McKinsey & Company found that decentralized models “with strong business ownership deliver greater impact and alignment with strategic goals.” [3]

Translation: If you want transformation, you need to embed it where the work happens. This isn't because centralization doesn't work--it does--but in the case of AI for Medical Affairs--it likely won't--and certainly not at its nascent stage.

WHAT THE OTHER SIDE MIGHT SAY AND WHY THAT STILL DOESN'T MAKE IT RIGHT

Some may argue that centralization brings benefits worth preserving. Let’s unpack that:

“Centralization promotes efficiency and eliminates duplication.”

Maybe. But efficiency isn’t the same as effectiveness. Innovation that’s misaligned with functional goals is just streamlined waste. Medical Affairs needs relevance more than it needs standardization.

“Medical Affairs hasn’t proven it can lead AI or digital—so let’s centralize it.”

That argument skips a step. Medical hasn’t led because it hasn’t been set up to lead. Starving a function and then blaming it for being small isn’t strategy—it’s mismanagement.

WHAT MEDICAL AFFAIRS ACTUALLY NEEDS

Medical Affairs doesn’t need to borrow someone else’s digital vision. It needs to build its own—and back it with commitment. That means:

• Hiring qualified digital and AI talent, not assigning roles to whoever’s available

• Embedding those roles fully inside Medical, with clear scope and reporting lines

• Setting its own innovation goals and metrics, not defaulting to Commercial KPIs

• Earning a seat at the enterprise table, not waiting for permission from it

Until that happens, centralization Medical AI Innovation will keep looking smart on an org chart—and keep struggling in reality.

#MedicalAffairs #AIinPharma, #PharmaInnovation, #DigitalTransformation #HealthTech #DecentralizeToScale #FieldMedical #PharmaLeadership #RethinkMedical #OwnTheRoadmap #PharmaAI #MedicalAI

RS CONSULTATIVE, LLC

*Narrative by me; editing by GenAI

References:

1. "The Last Days of Target Canada," Canadian Business, 2015 – https://canadianbusiness.com/ideas/target-canada-case-study/

2. “Why Innovation Labs Fail,” Harvard Business Review, 2020 – https://hbr.org/2020/01/why-innovation-labs-fail

3. “How to Make IT and Business ‘One,’” McKinsey & Company, 2021 – https://www.mckinsey.com/business-functions/mckinsey-digital/our-insights/how-to-make-it-and-business-one