Most AI programmes fail not because the technology isn’t good enough, but because ways of working don’t change. In pharma, the answer isn’t more tools. It’s a fundamental shift in how people learn and work alongside AI. If you want value, you need to treat AI fluency as a transformation capability.
AI in pharma is booming
The global AI-in-pharma market is approaching $2 billion and is forecast to exceed $16 billion within the decade.1 Many pharma companies are investing accordingly, with 95% reporting investment in AI capabilities.2
Across the value chain, from clinical development through to commercial launch, AI is transforming what’s possible, particularly in areas historically constrained by repetition, time pressure, and complex, fragmented data.
You can see the shift already:
AI has introduced a structural shift in how value is created across the pharma lifecycle. Yet Gartner predicts that over 40% of AI projects will be abandoned by 2027, because organisations haven’t built the ways of working needed to keep up with the technology.3
The myth: scaling AI is mainly a data and platform problem
There’s a persistent belief in pharma that AI scaling is primarily a data and infrastructure challenge, but our experience across global pharma tells a different story.
We’ve seen value stall when leaders and teams simply aren’t ready. They don’t yet have the confidence or the know-how to fold AI into everyday decisions. And in highly regulated environments, where getting it wrong means regulatory action, patient safety risks or lost public trust, that hesitation only deepens.
The real bottleneck is the gap between what AI enables and what your people are set up to do safely.
AI literacy versus AI fluency, and why the distinction matters
We’ve seen many organisations respond to this gap by launching AI literacy programmes through e-learning modules, prompting workshops or drop-in sessions. It’s a good start, but on it’s own, it won’t change how decisions get made.
Literacy is knowing what AI is. Fluency is knowing what to do with it. It’s the difference between understanding that a large language model can draft a regulatory submission and knowing when its output can be trusted, and how to govern its use in a Good Practice (GxP) environment.
In pharma, fluency carries a nuance that generic AI training can’t capture.
- Leadership fluency equips leaders to ask better questions about the decisions AI is now part of, such as risk.
- Practitioner fluency enables teams to apply AI within specific functions or workflows.
- Regulatory and ethical fluency helps employees understand the standards for interacting with AI-generated content or making AI-informed decisions.
When an organisation develops all three, something important shifts. AI stops being a tool and becomes a capability – one that enhances performance and unlocks possibilities that didn’t exist before.
The change management reality: fluency is built, not taught
Here’s what most AI strategies miss entirely. Fluency isn’t taught through learning alone, it’s built through change management.
These on-the-ground persistent misconceptions about behaviour change often derail AI adoption:
“Good communications alone will drive adoption."
They won’t. People can be perfectly aware of AI and still never use it. Awareness alone won’t deliver behaviour change.
“We just need to educate people on data and AI.”
Education creates knowledge, but it doesn’t create confidence, trust, habits or new ways of working. You can understand how a tool works and still not trust it enough to change how you make decisions.
“If people want to change, they just need to decide to do so.”
Behavioural change runs on reinforcement and the removal of friction, not on willpower. If the governance framework is too complex or confusing, people will route around AI tools no matter how well-trained they are.
“Changing attitudes will change behaviour.”
In fact, it’s often the reverse. When people have a positive experience with AI, their attitude shifts. Experience drives belief, not the other way around.
This is why the 70:20:10 model (McCall, Lombardo and Eichinger, 1996) remains so relevant. If we apply it to AI:
- 70% is built by doing: using AI tools in your actual work – drafting a regulatory submission with Copilot, pressure-testing a model’s output against your own judgement, and figuring out where it saves time and where it falls short.
- 20% comes from learning with others: sharing prompts that worked, debriefing on where a model hallucinated, building confidence through AI ambassador networks and peer-to-peer coaching.
- Only 10% comes from formal training: the e-learning modules and workshops that most organisations treat as the whole programme.
Yet most organisations invest as if this ratio were inverted.
For hands-on learning to succeed, two conditions are non-negotiable.
The first is psychological safety. People need to know that experimenting with AI – making mistakes, and asking for help – is not only encouraged, but expected. It’s about being bias-aware, responsible and keeping humans in the loop.
The second is a growth mindset – a culture that values curiosity over perfection and treats AI fluency as a journey, not a destination.
How leaders can enable change
As a leader, it’s important to recognise where your teams are today and provide the right conditions, as well as the right content, to help them. Build your own fluency so you can sponsor the right guardrails and give teams permission to learn through practice.
The way you talk about AI shapes how people feel about it. When leaders set the right tone teams understand the why, feel confident in the how, and become active participants in the do.
There are three key messages that we believe matter more than any training programme:
The prescription: treat AI fluency as transformation
If pharma leaders make one shift, make it this; stop treating AI fluency as a training problem and start treating it as a transformation problem.
In practice, that means:
Technology won’t close the gap between AI ambition and AI value on its own. Changing the way people think, learn and work will.
In pharma, where regulatory compliance and patient safety are at stake, it may be the most important capability investment of the decade.
Looking to close the gap between AI ambition and AI value in your organisation?
Get in touch to find out how we’re helping Life Sciences businesses successfully scale AI
