How AI will (finally) end the rep-driven pharma sales model

In 2011, I had the opportunity to work at Google, and my time there cemented a belief that technology would soon supplant the traditional pharmaceutical model of driving sales through live rep visits to doctors. It seemed like the ability to target physicians precisely and provide just the right information at the time it was needed, along with email and online video, would make sales reps obsolete.

I've rarely been so wrong about anything in my life.

What I didn't understand at the time was just how valuable an HCPs time really was. I assumed that a company which was able to swap out a live touchpoint with a digital one would save a lot of money and get a similar amount of attention. But because doctors are so busy, and there are so many marketers trying to get their attention, most non-personal tactics don't gain any traction with the majority of potential prescribers. So a rep visit, even if it looks inefficient on paper, is often the only way to guarantee a target customer actually hears your message and considers the brand at all. Sending a hundred emails or running banner ads on every site you think your customer might visit might, in practice, only reach a tiny percentage of your audience, which is overwhelmed and aggressively filtering out any information that they don't immediately need.

Before I speculate about a future where that changes, let me tell you what AI will NOT do in pharma marketing: we will not have AI chatbots that replace reps, giving highly personalized details to receptive physicians. That simply will not work, because doctors are not giving time to reps primarily because they find the reps deeply helpful (although some do, and some reps are truly amazing at what they bring to the table.) Mostly, reps get attention and advance marketing goals because they are a real human being asking for your help, and we generally want to treat other humans with care and respect when we can. Add to that that physicians do in fact want and need to learn about new drugs, and you can see why a live (or even virtual) interaction with a rep can demand their attention in a way that an AI detail simply won't.

The reason I see the role of the sales rep diminishing in the future (say in 5-10 years) is not because of AI's sales capabilities, but because of its drug discovery potential. How much it costs to develop an approved new drug varies, but a low-end estimate is over $300 million and most researchers who look at it holistically (including the cost of failed attempts and other expenses) say that it costs over $2 billion to get a drug approved.) With around 50 new drugs approved by the FDA in a typical year, that means new drug approvals are both very expensive and somewhat rare. A company that succeeds in running this gauntlet and getting a drug to market is going to spend aggressively on marketing to ensure that they capitalize on their success. However, there are already signs that AI can both help identify new drug candidates more effectively and that it can find additional uses for approved drugs. This could both lower the cost of development and increase the market potential for approved drugs. An easier development pathway and greater sales opportunities would make more drugs viable and increase the number of approvals. (If the FDA is able to effectively incorporate AI into its review process, that could also speed development timelines.)

Imagine, then, that we have a future where twice as many drugs are approved a year, and the drugs that are approved have more indications on average than before. This may dramatically increase competition (which is already intense) in areas like oncology and cardiometabolic disease, as well as grow the number of rare diseases which have approved treatments. This will further increase the learning demands on physicians--who are already struggling to keep up--and also put strain on sales reps trying to cover more doctors and speak knowledgeably about more conditions. If the number of reps grows to match the increased demand, physicians will feel even more overwhelmed and more of them will refuse to see reps at all. And if the number of reps stays more or less constant, they will be able to cover fewer of these new drugs.

How to relieve that gridlock? I foresee reps focusing only on the most high-value interactions, and AI technology filtering increasing amounts of other information to make it useful to physicians. Reps will focus on major drug launches and data releases, trying to ensure key physicians not only know the information but share it with their peer networks. But most marketing will need to be optimized via AI to reach the right doctor at the right moment. Teams will create communications that are designed to help AI filters and aggregators pull key claims, guidelines, prescribing information and other key information to physicians who need it. Imagine an oncologist gets a weekly digest from a trusted source like JAMA that includes study results most relevant to that physicians' current mix of patients. Marketers might pay for each time the algorithm serves their content to a physician, or for priority placement, or for a banner add that runs alongside their content when it appears. Basically, instead of omnichannel efforts to reach HCPs in lots of different places, we will have to optimize our ability to gain relevant attention in the increasingly small number of curated sources that physicians make time for.

This begs a key question, though: what if the AI intermediary isn't impressed by your drug or your data? In other words, what if you need to market more creatively? I would guess that "AI optimization" will emerge as a marketing specialty, just like SEO became a critical skill when Google was at its most dominant. But I also think tools like podcasts, debates and KOL interviews will become disruptive tools that allow alternative takes on a drug's value to reach a receptive audience. There will essentially be three types of drug brands: dominant brands that easily gain attention because their clinical benefits are widely appreciated, niche brands that spend minimally to fulfill a narrow use-case, and challenger brands who seek to upend current treatment approaches (and are willing to invest in marketing to do so) by raising the perceived value of benefits that are currently under-appreciated.

The odds are high that this technology will alter the dynamics of the drug industry in multiple ways over the next decade, and it is a safe bet that there will be some fundamental changes in the marketing model that most brands have relied on in recent years.