Most commercial pharma decisions are made on data that is weeks or months old. By the time a brand team sees the number, acts on it and pushes a change through, the window it was meant to catch has closed. Yesh Srinivasan, Chief AI Officer at Doceree, has spent the last year building a system that closes that gap, and on Day One of the Agentic Pharma Summit he takes the stage with Bill Veltre, Chief Media Officer at Deerfield Group, to argue about who should be making the call once the data is live: the human or the agent.
I sat down with Yesh on October 21 to talk through where commercial teams should start, what the word agentic means in practice, and what he has seen change inside the brands running his platform. The full conversation is above, about eighteen minutes. Here is what stayed with me.
Start with one brand
Yesh did not come up through pharma marketing. His PhD work was computer vision for early cancer detection, and most of the last decade was adtech. That outsider's view shapes his advice: start small, with one brand, and let the team build trust in the system before it is asked to trust it at scale.
"People start with one brand and ask what it can do, what questions they can ask about their campaigns and their measurement, and then extend it from there. Once they are comfortable with the reasoning it provides, the recommendations it provides, and the guardrails they are seeing, they make the changes it recommends and immediately see the change in performance. That moves the needle."
What agentic means
I get asked this question in nearly every conversation about the Summit, so I put it to Yesh. His answer was the plainest I have heard.
"Anything that acts with agency. Something that can get work done on your behalf. Things you can run in the background that are constantly monitoring something, summarizing the results, and saying: I have observed all of these signals and I am recommending that you take these actions. Then the human looks at the reasoning and approves the decision."
Thirty minutes, multiple dashboards, months-old data
The session title, Stop Reporting. Start Running, came from a line a customer gave Yesh's team during a Doceree summit: a typical morning is thirty minutes across multiple dashboards, looking at months-old data. Claims data runs weeks behind. Campaign analytics run months behind. By the time a decision gets made, it is usually a plan for next year's campaign.
"The gap between knowing and acting costs decision quality much more than it costs time. The title is about closing that gap."
Why now
Teams have talked about near real time pharma marketing for years. Yesh's view on what changed is short: the AI stopped being the hard part. Doceree has always collected signals, including direct connections to HCPs inside EHR platforms at the moment a diagnosis or a prescription is being made. What is new is a frontier model that can reason across those signals on a brand team's behalf.
"Now you have powerful models that can do a tremendous amount of reasoning and listen on your behalf, paired with these signals. You have the opportunity to go close to real time in how you build your campaigns, direct your campaigns and measure your campaigns."
The living media plan
I asked what happens to the traditional media plan, and to the agency relationship, once an agentic system is in the loop.
"It becomes a living, breathing plan. It is not one plan you create and then keep defending for a quarter. Our clinical intent signals are near real time, so you have an opportunity to change the doctor's mind while the decision window is still open. You modify the plan, measure in real time, see whether the results are directionally correct, and the outcomes are infinitely better."
Brand teams, in his telling, move from building and defending plans to directing them: where to put budget, which creative to change, which states should get in-person reps this month. The data scientists are still there. They just get pointed at a specific question instead of "go figure out why my campaign is underperforming."
The biggest misconception
Yesh's answer here is the one I would want every commercial leader in the room to hear before November 9.
"The biggest misconception about AI agents in pharma marketing, and about AI in general, is that they are built to replace the people running the brand. The model is advisory. It reads, ranks, proposes, and every move waits on approval from the human. The shift is from dashboards that report to a system that recommends the next best action."
Why be in the room
Yesh and Bill Veltre lead Stop Reporting. Start Running: When Agents Take Over Pharma's Commercial Model on Day One, Monday, November 9, at 11:30 AM. In Yesh's words: "You get to watch a chief media officer and a chief AI officer try to agree on who should be making the decisions, humans or agents. We have thirty minutes. I am optimistic. If you run a brand, you will leave knowing how close this already is and what to do about it."
Doceree is a Founding Partner of the Agentic Pharma Summit. Seats are limited; register here.
Agentic Pharma Insights is written for pharma commercial leaders working through what agentic AI actually means for marketing, sales, patient experience, and access. See the 2026 Summit agenda.