This article is part of solli’s September series exploring The Effective Use of AI in Pharma Media.
You are about to start the biggest race of your career. Strangers are rooting for you, with their hopes and dreams riding on your success. A strong start almost guarantees you the win.
I’m describing not an Olympic sprinter, but the launch of a new therapy: one that is first in its class and will offer years of well-being to patients and peace of mind to their families.
You and your team are ready to build your plan and execute—it’s go time. So, when an organizational mandate arrives to infuse your launch with artificial intelligence, it is your job to convert that mandate from a distraction into an asset and advantage over your competition.
In the U.S., where the typical patent life of a branded drug is seven to 10 years post-approval, there is no time to waste in getting patients on therapy and monetizing the work that was necessary to bring an innovative medicine to market. However, there is a very real tension between those leading a brand and those in the departments they depend on to execute. This is nothing new: The idea of what is good for the brand versus what is good for the portfolio has always existed. This is a good thing; this is how we get to innovation and improvement on both sides.
AI has sharpened this tension, as a moving target and a source of pressure to implement. Here are three considerations to take advantage of what is possible now and escape the distraction that can come from embedding AI—both for yourself and your extended team.
It is critical to remind everyone what it is you are trying to do—especially those ordering the use of AI. Most mandates and experiments are set aside if you can move a prescriber or patient to therapy faster.
Ultimately, what you choose to do should compress the time to therapy and ought not cause confusion or apply undue pressure to those you work with. Some well-documented examples of AI applications in pharma media that fit the bill include segmentation, media optimizations and content variants.
With 95% of enterprise generative AI pilots delivering no measurable ROI, according to a 2025 MIT report, you’ll want to go where there is high confidence of impact. This is not to say you should not experiment; rather, defer those experiments.
Keep it simple: Can you isolate the AI impact to a single activity or capability? If you are considering a change that goes across disciplines, the risk is that you will be more focused on fixing a process than on the growth of a brand. While the people who come after you will appreciate a better process, this is not going to help you cross the finish line first. Save this for outside of launch.
I hear it consistently: Clients are expecting rate-card reductions from AI. This thinking has proven to be premature, however, as AI-assisted review has not consistently proven reliable across agency and client workflows.
With lower cost per asset, there will be a commensurate reduction in agency review and shepherding to client for review and MLR—which is critical at this early stage of automation. Disrupting your MLR workflow hurts productivity and has a compounding risk of reducing your reviewers’ confidence in future submissions.
Then there is the Jevons Paradox. Even if you don’t know it, you recognize it. When a resource becomes cheaper and faster to make, usage increases. With desktop publishing, for example, when newsletters became easier to produce, more of them appeared. The same dynamic is already emerging with the volume of AI-generated content variants. Lower cost per asset means more assets, which means more to review and more to manage—by the same number of people or even fewer.
Building your own AI capability was the right call a few years ago. The vendor market didn’t exist beyond management consultancy prompt rubrics, and we were all learning by doing. That has now changed.
In 2025, MIT found that externally partnered AI deployments reached production at twice the rate as internal-only builds, 67% compared to 33%. But that doesn’t necessarily mean handing it over to outsiders is now always the right choice. The marketing AI marketplace is still nascent. Vendors’ knowledge base and experience may not be that far ahead of your own. So, as you make decisions, ask: Does the specialist market now beat what you can do? Do they have the experience that you need for the job?
Importantly, you will want to have your criteria for decisions ready, because you may already have sunk costs and internal pride to evaluate against. And after you decide, set a date to revisit your decision. The AI vendor market will look different in 12 months, and your answer might, too.
Work through these three points, and what once felt like an organizational headwind is now the tailwind that will carry you to your best race. The starter pistol is up. Your heels are up at the starting block. Patients are waiting at the finish line.
For more on The Effective Use of AI in Pharma Media, click here.
This guest commentary piece was written by Ben Versh, director of content strategy at Pfizer Inc.