Senior Enterprise Leader, R&D
Poonam Rihel has watched a decade of digital transformation initiatives in pharma produce more overhead than outcomes. Her diagnosis: organizations keep layering technology on top of broken processes instead of fixing the foundation first. Now, as agentic AI moves into regulatory writing, she argues the industry faces the same temptation—and the same risk.
Clinical reporting sits at one of the most consequential intersections in drug development: where science meets execution, and execution meets regulators. Get it wrong, and submissions stall. Get it right, and you compress timelines across the entire portfolio. Poonam Rihel has spent her career leading across the R&D value chain, and now with focus on regulatory and medical writing to redesign not just the documents they produce, but the organizational logic that produces them. As AI begins to reshape what “writing” even means in a regulated environment, her perspective on what actually drives transformation—and what merely mimics it—carries real weight for anyone responsible for clinical operations, regulatory strategy, or the people who sit between data and submission.
Moe: Why did clinical reporting draw you in specifically?
Poonam Rihel: Clinical reporting is where regulatory writers, medical writers, and scientific writers sit at the interface between science, execution, and regulators. It’s one of the most interesting places to be in the pharma value chain—early portfolio, late portfolio, everything that happens along the way. I was lucky to be met with a talented and innovative focused team, committed writers who could see there was a different way of doing this, and who knew their craft well enough to understand that structure, standardization, and templates were the first step to actually doing something different.
Moe: Why rebuild the process before layering in digital tools, rather than automating existing workflows?
Poonam: I’ve had the fortunate and unfortunate experience of trying it both ways. Across the last decade of digitalization, many of us in pharma have gone in with the big dream—put in the system, and it should solve everything. What you actually do is increase overhead, create pseudo-work that wasn’t there before, because you haven’t reimagined what the layer underneath looks like. What is the process? Who is responsible? You move your existing flow into a system that might just give you different problems. That experience led us to recognize that there’s something about standardization, about scientific intent, about defining the process and who does what, that actually gets you end-to-end faster than just layering on a digital solution.
Moe: How do you respond when someone insists their study report is the exception to the rule?
Poonam: It is not the exception. There are of course variations, and I think there’s sometimes a misconception that standardization means locked. Standardization means you leave space for innovation within a frame that is focused and defined. Things can, and do change—we work in project organizations, we don’t know what data is going to come out, we need to be flexible and agile. But being clear about the guardrails, not allowing change for change’s sake that doesn’t add scientific clarity or execution speed. It is possible to reverse-engineer and compare if changes made across projects, submissions, and documents had measurable impact. We’ve had the most success pointing medical, clinical, and regulatory teams toward the sections where impact is greatest, and now we move into amplifying that across the value chain on more documents and content.
“Standardization means you leave space for innovation within a frame that is focused and defined.”
Moe: How does change management actually work in large pharma organizations—from planning through stakeholder buy-in to overcoming resistance?
Poonam: We’ve gotten our battle scars along the way. There’s definitely something to “build it and show them”—the proof is in the pudding if you can re-engineer a process so it completely changes how people interact with it. That’s a start, but it’s definitely not the end. Then comes a lot of communication, a lot of working with different stakeholders, a lot of standing your ground, a lot of convincing. It’s a massive task. But we do see that when you invest the time, people come along with you on the journey. Making it an enterprise change rather than a forced change means that you get traction and buy-in along the way. Both approaches can work, but what you want is adaptability to change, and bringing people along so they can see how this benefits them and a solid value proposition. It’s been an important part of our innovation focus.
Moe: How does an organization hold on to deep expertise before it walks out the door as AI takes over more of the work?
Poonam: We’re being very clear about this. We work in a regulated environment, which means there always has to be a human who decides—humans are in control, and AI is in the loop, not the other way around. We’re specific about where we want to use AI or automated processes, and about which competencies we need to amplify to maintain a good balance—not losing interpretation skills, not losing facilitation, not losing decision-making. Right now, nothing gets out the door without us having made the decision and pressed the button, and that’s how we continue to look at it.
For us, it’s not about fewer writers—it’s about amplifying individual impact. The amount someone can do today, we hope and expect that we walk into a future where they can do more: other documents, parts of the value chain they never touched before. Automation and AI can get us a long way, but tone of voice, strategic thinking, scientific red threads, narratives—there’s still something there for a while at least. I’m sure AI will surprise us at some point in areas we think are very strong to hold on to. So we have it as a focus to build capabilities and competencies in those areas alongside the AI piece.
Moe: How does a defensible AI audit trail for regulatory submissions actually need to be structured?
Poonam: It’s an absolute work in progress and a focus for many of us across pharma. What does good look like in this space? It needs to be defensible and auditable. We need to break the process and the thinking down—where are we signing something off, where are we making a decision, where are we pulling data and information from, what context has been applied, which LLMs are we using. Most authoring models aim for creating a draft, and that leaves accountability with the writer to have executed their “human” process with the same diligence as without AI. The sweet spot is somewhere in the middle.
Moe: Why does human expertise in medical and regulatory writing still matter, and where does it matter most?
Poonam: The first and most important thing is that their expertise matters. A writer who can see a scientific red thread, pull out strategic nuances, tell a story, communicate—those skills still matter. The question is how to apply them going forward: how to be the voice that says, this skill applies right here. It’s about working together with AI and technology, not having AI and technology work against you. The more writers become familiar with the ups and the downs, the pros and the cons, the more they can be a voice that shapes where this goes. The skill set behind writing—facilitation, scientific understanding, reasoning—I don’t see that going away. Even if AI does it, it doesn’t mean a human shouldn’t do it as well. The challenge is how to build that capability when AI is doing more and more of it.
“Even if AI does it, it doesn’t mean a human shouldn’t do it as well. The challenge is how to build that capability when AI is doing more and more of it.”
Poonam Rihel is a Snr Enterprise leader in R&D and is currently driving digital transformation in clinical reporting, regulatory writing operations, and leveraging AI integration in drug development.
Moe Alsumidaie is Chief Editor of The Clinical Trial Vanguard. Moe holds decades of experience in the clinical trials industry. Moe also serves as Head of Research at CliniBiz and Chief Data Scientist at Annex Clinical Corporation.
