At a healthcare media event in New York City late last month, a panel discussion about broken operating models veered, as so many of these sessions have since the arrival of ChatGPT in November 2022, into a panel discussion about AI. The panelists noted that AI could read a brain MRI more accurately than most radiologists, help researchers locate candidates for their under-populated clinical trials and, of course, automate a great majority of what sales reps do on a day-to-day basis. All of these statements were accepted as established fact by panelists and audience members alike.

The first two may still be open to debate. The third one isn’t, or at least it shouldn’t be.

So once more, with feeling: AI is not going to replace the pharma sales rep. AI doesn’t allow for effective two-way exchanges of information. It can’t reliably navigate thorny compliance questions. It can’t read a room.

Over the last half-year or so, a relative consensus has emerged around this line of thinking. Namely, that AI can automate the parts of the job that surround the rep/HCP conversation – transcribing notes, surfacing relevant content, identifying next-best actions – but not the conversation itself, nor the relationship underlying it. That it took the industry three years to arrive at this conclusion speaks to the extent to which AI has hijacked the imagination of leaders across the healthcare spectrum.

“I think the right way to think about AI is as an assistant to the rep, not a replacement,” says Jonathan Bentley, SVP, sales, operations and training at Phathom Pharmaceuticals. “Where the human piece still really matters is in judgment, context and trust… Every interaction is a little different.”

Tyler Redelico, director, omnichannel and content strategy, US Medical Affairs, at AstraZeneca, goes one step further. “Today? I don’t think we’re at ‘fully automatable’ with anything,” he explains. “But there’s a lot of opportunity to better equip our field teams with insights so that they can execute better. And we can get more precise about which HCPs to target at the right time.”

That precision matters, especially at a time when HCP satisfaction with pharma engagement has tumbled. A Deloitte study conducted in March 2025 revealed that a mere 28% of HCPs feel that pharma’s customer engagement strategies meet their needs effectively. Meanwhile, 34% said that sales reps convey scientific narratives imprecisely and fail to adequately tailor their messaging. 

Should that be interpreted as a sign that the venerable pharma field force is starting to outlive its utility? The most distilled version of the efficiency argument for reducing the size of field teams goes something like this: Few HCPs value the information reps share as much as they once did. If an HCP needs product data or clinical trial results or dosing guidance, they’re easy enough to find elsewhere without waiting for a rep to hand-deliver it. 

But that doesn’t take into account institutional knowledge or situational awareness. Consider, for example, a rep who has been dropping in on the same practice for a half-decade. This rep is used to navigating access issues that, in the past, have proven a thorn in the practice’s side. For conditions treatable with multiple competing medications, this rep is able to provide an informed level of support around billing, savings and patient education.

Good luck replacing all of that with a few prompts. As Bentley puts it, “There’s a difference between knowing what to do and actually being able to do it. It’s less about just delivering information and more about understanding what information the HCP actually needs, asking the right questions and making the interaction more relevant to them.”

Phathom, Bentley notes, has been using AI to increase rep efficiency around pre-visit prep work, post-visit follow-up and the identification of market trends. “[The rep] can only be in so many places in a day, so increasing efficiency is critical,” he adds.

The company has already seen genuine progress: Even small, AI-fueled improvements in preparing reps for HCP conversations have meaningfully enhanced the quality of those conversations, Bentley reports. “[AI] doesn’t replace the conversation. It just elevates it.”

That thinking, most observers agree, represents another emerging consensus. To fully realize it in real-world settings, however, will require comprehensive company-by-company education – or maybe re-education – efforts. 

In discussions about the adoption of AI, it’s often underplayed that many pharma reps have been doing their jobs (and presumably doing them well) for quite some time. While tech-minded leaders may want members of the field force to get on board with the newest and flashiest digital tchotchkes, they can’t glaze over the change-management piece of the exercise.

“Some reps and doctors, it’s fair to call them ‘technology-challenged,’” notes Irvin Carlin, formerly associate director, digital marketing, HCP and consumer, at Currax Pharmaceuticals. “That’s not a knock on them. It just means that, if you’re trying to get them to realize whatever it is you want AI to do, you better make sure they’re trained on a workflow process. There needs to be an overarching strategy and also a change-management/workflow strategy.”

There must be a similar emphasis on internal leadership and training. Most pharma companies have teams, usually composed of former reps, devoted to educating members of the field force. Perhaps those individuals can add AI integration to their topical mix, or maybe an outside consultant might be tasked with codifying organizational practices around AI.

“Companies need to identify this as an area of investment,” Carlin says. “They have to make a case for the value it brings, because already you’re seeing plenty of skepticism around AI in general.”

That skepticism is finally being confronted head-on. Reps often reminisce about when they were first armed with iPads, warmly recalling the wow factor that came with using the then-groundbreaking technology in HCP detailing. They’re a bit less sanguine when they discuss the pandemic-era shift to Zoom meetings – the impersonality of which, they believe, impaired their ability to educate and forge meaningful relationships with HCPs – and even less so when talking about the current AI revolution.

So even as AI automates a slew of responsibilities that have traditionally fallen on pharma reps, they can remain indispensable by being recognizably human in every exchange. Redelico, who initially joined AstraZeneca as a medical science liaison, believes that reps and MSLs alike can better serve HCPs by adding a personalized touch to these interactions.

“When I was an MSL, I developed great relationships with HCPs by adding color to the messaging I was trained on and the slide decks I was presenting,” he says. “Some slides need commentary and some exploratory analyses need caution, and approaching this with nuance is critical for rapport and trust.”

As for what lies ahead for field teams, it would be a huge upset if human/AI partnerships aren’t accepted as standard practice within the next half-decade. Bentley anticipates a continued shift from push to pull, with HCPs tracking down more information on their own terms. “Then AI creates the opportunity to meet them in those moments with timely, relevant information, and then connect them to a person when deeper engagement is needed,” he explains.

Indeed, the most likely outcome isn’t a rep-free commercial model; it’s a more finely calibrated one, with more AI-assisted prep work and human-led problem-solving and fewer low-value interruptions. It will then be incumbent on reps (and the leaders training and directing them) to balance the “old” parts of their job with the new ones.

“It’s going to take curiosity and a willingness from leaders to try to do things differently. And there will have to be an acceptance that maybe they will fail at first,” Carlin says.

Bentley agrees, adding, “At the end of the day, people don’t like to be sold to. They like to make decisions. The rep’s role is helping create the environment for that.”

Has your organization leveraged AI to maximize the efficiency of your field force? If so, how? Drop us a note at hello@kinara.co, join the conversation on X (@KinaraBio) and subscribe on the website to receive Kinara content.