At the Reuters Events Pharma conference in Barcelona this spring, a group of senior professionals from commercial, medical, and market access gathered for an 80-minute session with an unusually direct premise: sit down, pick a problem, and design an AI-powered solution for it before the session ends. The problem domain was omnichannel product launch – a discipline the session opened by characterising, with some candour, as one where industry performance routinely falls short of what companies expect.
Why launch execution stalls
The reasons for that shortfall, as Ivanka Vankova (Head of Life Science Sub-Practice) and Krystian Jablonski (Director, Data and AI Strategic Solutions) laid them out on behalf of Lingaro, are well understood, if rarely stated plainly in a conference room.
According to them, launch challenges include the following:
• Review cycles run long, particularly for materials requiring medical, legal, and regulatory sign-off.
• HCP engagement runs on fragmented signals. Field teams, MSLs, and omnichannel leads often operate from different data sources, making it difficult to sequence the right action, channel, and message at the right moment.
• Global brand materials require local adaptation for language, compliance, and market context, but the process for doing that at speed and at scale is rarely in place before launch day.
• Clinical, commercial, regulatory, and patient data each sit in separate systems with their own definitions and quality standards, so the insights that should be driving decisions during launch are often delayed, inconsistent, or simply not trusted.
The gap between plan and execution accumulates not so much from a single failure but from an accumulation of small disconnections across the whole operation. Each seems innocent in isolation, but in concert, the impact can be devastating.
Data first before AI
Lingaro’s answer to that set of problems begins, by design, before any AI is introduced. Wiktor Fido, the company’s general manager for European markets, framed the organisation’s position at the outset: data should come first.
This is something most organisations miss. In Lingaro’s recent research, The State of AI Readiness in Pharma, they found that 67% report fragmented or only partially reliable data across their core domains. That means two in three pharma organisations are building AI initiatives on a data foundation they do not fully trust. Every model, dashboard, or AI assistant deployed on top of that foundation inherits the same risk.
The argument goes like this: agentic AI – systems capable of taking sequences of actions without step-by-step human instruction – is only as reliable as the data it operates on. Organisations reaching for AI before their data foundations are sound are unlikely to see returns worth measuring.
Solution concepts straight from pharma leaders
Krystian Jablonski (Director, Data and AI Strategic Solutions), who facilitated the session, described a concept he called an “agent factory”: a framework giving business users the ability to build their own agents within their own company data, without routing that work through IT.
The tool that participants worked with was the Solutioning Cockpit, Lingaro's proprietary conversational AI environment. Users state a challenge, then the system decomposes it by asking clarifying questions. The two then work toward a designed solution, producing a downloadable output covering the problem statement, proposed solution, time to delivery, estimated costs, and a visual representation of the finished design.
Five use cases were offered to participants as starting points, each targeting a distinct stage of the launch process. Participants were encouraged to treat any of these as starting points rather than fixed briefs and to bring their own problems if none of them fit. The use cases shared were:
1. An MLR pre-screening tool designed to reduce the time spent on review cycles during launch preparation.
2. An alignment monitor, a solution to the visibility problem between teams: giving each function a view of where the others are and where progress has stalled.
3. The concept of next best action, usually applied in a narrow commercial context, and extended it to the full range of HCP interactions at launch.
4. An evidence synthesis tool to reduce the time required to work across large bodies of clinical material.
5. A localisation capability aimed at the stage where global content must be adapted for individual markets, a process that currently adds another round of sign-off and further delay.
From these use cases, participants built a total of nine solution concepts. For example, the AI Rep Readiness Coach applied AI-driven coaching to improve field readiness and adaptability during launches. There was also the MedSynth AI, which tackled the evidence burden directly, accelerating literature review through AI-assisted synthesis while keeping traceability intact. The winning concept was the CLIP Launch Intelligence Platform, recognised for directly addressing fragmented launch coordination and for the clean separation of medical and commercial data within a single intelligence layer. That distinction matters in pharma, where the consequences of data crossing the wrong boundary are not abstract.
One practical note ran through the session's facilitation: that AI, left to its own devices, will continue to generate clarifying questions indefinitely. Lingaro’s expert team was direct about the countermove required: “AI is ultra-curious; it’s going to ask you questions. But we don’t want more questions. We [want] solutions." Pushing the tool toward a first draft rather than an indefinite refinement loop proved as important as any technical consideration.
What leaders and AI can do, together
The session’s organising conviction was stated plainly in its closing minutes: "AI is not here to replace us. Our work can be much accelerated with the use of AI, and the outputs of our intelligence can be amplified massively." The hesitation around AI adoption in pharma is real. Based on Lingaro's findings, adoption and change management is the single biggest barrier to AI success, cited by 31.6% of respondents as their top obstacle. Only 7.1% reported having AI tools fully embedded in their daily workflows.
Fear of replacement is a common undercurrent in any conversation about AI at work. But Lingaro’s position is clear: AI is not a substitute for the expertise already in the room. It is a mechanism for acting on that expertise faster and at greater scale, provided the data it runs on is ready.
The design of the five use cases reflected that position directly. Each one targeted a stage of the launch process where the delay was measurable, its source known, and the human expertise required to describe a solution already existed in the room. In Lingaro’s model, what AI contributed was the speed and scale to act on that expertise, provided the data was in order before the agent was switched on.
That is ultimately what the hackathon demonstrated: that pharma organisations already have the knowledge to break through launch barriers. The right AI solutions, built on ready data, give them the means to act on it.
For more insights on commercial pharma, including latest articles, interviews and more – click here.
