In this solo episode of AI For Pharma Growth, Dr Andree Bates explores why many pharma teams do not have an AI problem at all. They have a capability problem.

Dr Andree starts with a simple question: when was your team last properly trained on AI for their specific role? Not when they were given access to tools, licences or a generic use policy, but when they were trained to use AI effectively, safely and compliantly in their actual workflow.

The episode challenges the usual explanations for disappointing AI results: the model was not good enough, the vendor was wrong, the data was not ready, or the organisation resisted change. In many cases, the tools work, the pilots are useful and the training lands. But the working knowledge needed to use AI well is uneven, fragile and decays over time.

Dr Andree explains why this matters so much in pharma. High-value AI work is often judgement-led: medical information responses, payer materials, safety narratives, regulatory documents and MLR-compatible content. AI can support these tasks, but only when users can tell the difference between a strong draft and a merely plausible one.

She also discusses the research behind skill decay, including why cognitive and accuracy-dependent skills fade faster than simple speed-based or physical skills. That is especially important in pharma, where the cost of a confident but wrong output can become a compliance, regulatory or patient safety issue.

The key message is clear: AI capability is not something you achieve once. It has to be maintained. The functions that lead in AI will not simply be the ones with the most licences or training events. They will be the ones that treat capability as something with a rate of decay and build systems to keep it current.

Topics Covered

Why AI underperformance is often a capability problem
The difference between access, policy and real training
Why confident AI use varies across teams
AI in judgement-led pharma workflows
Skill decay and why 90 days matters
Why high-value AI workflows are often forgotten fastest
The risk of outdated working knowledge
Why training is ignition, not maintenance
The limits of AI champions and internal portals
Three questions to ask your function this week

Eularis helps pharma and biotech leaders turn AI activity into board-defensible governed strategy and measurable commercial outcomes.

If your CFO asked tomorrow for the projected return of each major AI initiative - by year, across three years, with explicit adoption, operating cost and redeployment assumptions - could you produce an answer that survives scrutiny?

And if you could: would you know which of those initiatives most moves the company toward the outcomes it's exposed on over the next three years? Those are two different questions, and most organisations can't answer either. A strong initiative-level ROI tells you a project is defensible. It doesn't tell you it belongs among your top five. Capital spent on a second-order opportunity is capital no longer available for a first-order one — and no amount of downstream rigour recovers value that was never strategically prioritised.

The Eularis AI Strategic Blueprint models both levels: a financial case for every prioritised initiative, and a rigorously modelled ranking of which ones create the most material value against your commercial objectives — then sequences them by dependency rather than enthusiasm, with governance designed for pharma's regulatory reality. See what a board-defensible AI strategy contains → eularis.com/ai-strategic-blueprint-for-pharma

About the Podcast
AI For Pharma Growth is the podcast from Dr Andree Bates, helping pharma, biotech and healthcare organisations understand how AI-based technologies can save time, grow brands and improve company results. The show demystifies AI for biopharma leaders, from start-up biotech through to Big Pharma.

Dr. Andree Bates LinkedIn | Facebook | X

E231: The Diagnostic Room: You didn't have an AI problem. You had a capability problem.

AI For Pharma Growth

E231: The Diagnostic Room: You didn't have an AI problem. You had a capability problem.