The PMSA Annual Conference is one of the anchor events for analytics professionals in pharma and biotech. This year it ran in New Orleans, May 3 through 6, under the theme Convergence of Data, Talent and AI. C&F hosted the Lunch & leadership roundtable. Three weeks on, the takeaways the team — Lukasz Drejka, Daniel Fracas, Kylene Merritt, Marcin Ludzia, and Kamil Skibiak — brought home are sharper than the average post-event recap and worth sharing in one place.
A roundtable built on real client work
On Monday, May 4, Daniel Fracas hosted the C&F luncheon roundtable, “From Use Cases to Velocity: How Data, Talent, and AI Deliver Real Pharma Impact.” Marcin Ludzia and Kylene Merritt joined Dan in moderating. The structure was deliberately simple: three real applications from Fortune 500 client work, then an open conversation about the patterns that connect them.
The three cases were an AI-powered bot for vaccine market insights (AI VaxBox), a Return on Data Dashboard that quantifies the actual business value of third-party data assets, and a three-tier Intelligent Care-Gap Signal Service that surfaces actionable opportunities from unstructured data. The common thread was an AI-driven data engineering approach designed to accelerate data completeness and velocity, lower cost, and improve quality across the lifecycle.
Kylene, attending PMSA for the first time, captured what made the format work:
“I left energized and more convinced than ever about the value of bringing people together. It was a great roundtable discussion around real-world applications and what actually moves the needle.”
All roundtable participants also took part in a tablet giveaway.
What we saw on the floor
Marcin Ludzia walked the conference floor with an analyst’s eye, and his read on the market is worth sharing in some detail.
The pattern he kept coming back to was the convergence of vendors around what he calls Decision Intelligence, even when none of them used the term:
“Almost every vendor on the floor was selling what I would call a Decision Intelligence solution. None of them used that term. They each picked a narrow use case: pre-call planning, agentic forecasting, conversational BI, market access automation. Under the hood, same architecture every time: pull signals from multiple sources, generate a recommendation, put it in front of someone who needs to act.”
Insight generation, in other words, is being automated at speed. How quickly can a rep get a pre-call brief? How many sources can the system synthesize? Those are largely solved problems now. The harder conversation, Marcin noted, is connecting that automated output to how decisions actually get made inside large pharma organizations, and that conversation is just starting.
LLM sessions kept circling the same architectural problem: general-purpose models struggle with domain-specific commercial data. The solutions gaining real traction were semantic layers, Bridge Data Layers, and knowledge graphs, all of them purpose-built context layers that ground AI in pharma-specific logic. Watching that pattern solidify across multiple vendors was one of the more telling observations of the week.
The observation Marcin found most telling pointed to where the conversation goes next:
“Data quality came up only a handful of times, but every time it did, the room shifted.”
He read that as a sign of real appetite for a conversation the market is just getting ready to have. Insight generation has raced ahead; the data foundations that feed it are what comes into focus next, and in his view that gap closes soon.
The bottleneck has moved
Marcin’s read on the data quality gap did not come from his sessions alone. The same theme surfaces from a different angle.
Kamil Skibiak put it plainly:
“AI capabilities have accelerated fast and they’re no longer the bottleneck. What is holding things back now is how data actually gets delivered, managed and organized.”
Kamil pointed to the specific shape of the problem: data inconsistency across teams, missing foundations, pipelines that fall behind, and governance and semantics that get deferred until they become a blocker. The destination, in his framing, is Decision Intelligence that produces measurable business value. Getting there takes foundations, an execution model, and the discipline to apply both to real use cases.
Kylene saw the same thing from the user side:
“Success comes down to the 80% below the surface and being strategic with your implementations. That means really knowing your users, building strong foundations, and staying intentional to cut down on the slop.”
Dan, who hosted the table, came back to a related theme afterward. The organizations getting real value from AI, in his read, are the ones treating their data as an engineered asset, and that point came up in conversation after conversation across the week.
Looking ahead
The PMSA 2026 theme, Convergence of Data, Talent and AI, framed the week well, but the synthesis we are taking back to client conversations is simpler than the headline. AI capability is no longer the hard part, but the data underneath is: how it is engineered, governed, made consistent, and made fast enough to act on. That is where the work happens, and it is the work C&F focuses on with our clients.
Thanks to Dan Fracas, Marcin Ludzia, Kylene Merritt, Kamil Skibiak, and Lukasz Drejka for representing C&F in New Orleans, and to everyone who stopped by the lunch and the booth to make this a useful week.