In a 2025 LinkedIn poll reported by the Medical Science Liaison Society, 72% of 806 respondents selected Medical Information as the Medical Affairs role most likely to experience significant reductions because of AI.

These were expectations from a self-selected poll, not evidence of actual workforce reductions.

Still, the reasoning is understandable. Medical Information involves literature searches, scientific response development, and answering recurring questions. AI is increasingly capable of supporting these activities.

But the poll answers a staffing question, not a value question.

Reducing the work required to answer a scientific question doesn’t necessarily reduce the value of the function providing that support.

And it certainly doesn’t reduce the value of understanding why the question was asked.

A JAMA study published October 15 examined how U.S. clinicians used OpenEvidence, an AI-enabled clinical information platform, between January 2024 and June 2026. During that period, 952,008 U.S. clinicians submitted 271 million queries.

Among physician queries analyzed:

OpenEvidence provided the data, and several co-authors work for the company. The researchers used AI to classify a sample of 38.8 million queries. The study doesn’t establish whether these tools improve clinical outcomes, when questions arise during clinical workflows, or whether they replace traditional scientific interactions.

The study demonstrates substantial use of readily accessible clinical information. It also raises a question about timing. A reliable response is most useful while the scientific or clinical question remains open. A thorough discussion several weeks later may provide greater depth, but it may not address the original need at the time it arose.

That brings us back to Medical Information. Unlike traditional planned Field Medical engagement, Medical Information is largely demand-driven. HCPs approach the organization with questions they want answered.

Some are straightforward. Others require more interpretation, discussion, or follow-up. But these inquiries have another potential value: they provide visibility into what HCPs are trying to understand.

There’s a complication, though. As clinicians increasingly turn to independent AI platforms, some questions may never reach the manufacturer.

Medical Information could become more efficient at responding to inquiries while manufacturers lose visibility into some of the questions HCPs are asking.

Neither Medical Information inquiries nor traditional KOL engagement provides a complete picture of the market. That makes connecting different sources of scientific information increasingly important.

This isn't an entirely new idea.

A 2025 phactMI benchmarking study found that 30 of 34 responding Medical Information departments already generated insights from inquiries.

Among departments responding to additional questions:

The study also found that 25 of 30 departments developing insights relied on manual review of inquiry databases.

Many organizations already recognize Medical Information’s broader contribution.

The harder question is whether those insights are consistently connected with what Field Medical is learning, and whether that combined understanding influences scientific priorities and organizational decisions.

Consider a recurring question about treatment sequencing that appears in Medical Information inquiries but hasn't surfaced through established Field Medical relationships.

Are HCPs struggling because the evidence is incomplete? Because existing information is difficult to find? Or because applying the evidence in clinical practice is more complicated than anticipated?

Those possibilities require different responses.

Medical Information can identify the recurring questions. Field Medical may provide additional context through scientific exchange. Together, those observations could help determine whether further investigation, education, or evidence generation is warranted.

The point isn’t to collect more insights. It’s to understand what they mean and what, if anything, the organization should do differently.

There's another advantage we may be overlooking: scalability.

Medical Information already has a delivery model designed to respond to inquiries from a broad HCP population without requiring established relationships. AI could extend that capability further.

But not every scientific need ends with an answer. Some require deeper discussion, continued follow-up, or sustained scientific engagement. That’s where Medical Information and Field Medical can complement each other.

The opportunity isn’t simply to make Medical Information more efficient or Field Medical more accessible. It’s to connect these capabilities so that scientific support follows the need rather than the organizational structure.

Manufacturers are making different forms of scientific support more directly accessible to HCPs. Consider these publicly available examples.

These aren't necessarily new capabilities. Medical Information has handled inquiries and facilitated scientific follow-ups for years.

What varies is how manufacturers organize access. Some provide separate pathways for Medical Information and Field Medical. Others allow HCPs to choose whom they'd like to hear from or how they'd like to receive a response.

Takeda's U.S. portal illustrates how both functions can be accessed through one interface.

But the interface doesn't tell us what happens behind the scenes. Are inquiries appropriately routed, resolved, and used to identify broader scientific needs?

Making scientific expertise easier to access isn't the same as making scientific support more useful.

Before deciding how AI should change Medical Affairs resources, I'd start with a simpler exercise.

Choose one recurring HCP need and trace how your organization responds to it.

Then examine the staffing, technology, workflows, and field deployment supporting that process.

The fifth question is particularly important. Medical Information could respond efficiently to thousands of inquiries without those questions meaningfully informing scientific strategy.

Field Medical could conduct hundreds of valuable scientific interactions while important needs outside its established stakeholder network remain undetected.

Both functions may be performing well against their existing measures, yet the organization may still miss important scientific needs or fail to act on what it has learned.

If we were designing scientific support around what HCPs actually need, what would we resource differently?

Pull the five most frequent Medical Information inquiry themes from last quarter. Ask your Field Medical leads which ones they're hearing in the field.

Where do the findings overlap, where do they differ, and what might explain those differences?

You may uncover a scientific need worth investigating or discover that your teams are seeing different parts of the same clinical environment.

Reply and share what you found, or what you expect to find.