Bina Patel of Amgen and Tom Ambrosole of ArisGlobal discuss how PV managers must lead their teams from the front.
A 2026 Global Labor Market study by Gartner of over 12,000 employees and managers revealed that only a fifth of leaders currently believe their workforce to be “AI ready”[1]. Frustratingly, it is often the operational teams and roles (those with the most to gain from smart automation) that feel least equipped and supported. This is something that now needs to change.
In most cases, the problem is one of uncertainty and fear. The teams in scope may not have had the opportunity to become comfortable with using AI in the context of their work, and may be concerned that they do not know how to minimize any risks — of using a tool incorrectly or inappropriately, for instance. To use AI effectively they must first feel confident that they are doing it right, achieved via appropriate change management.
As part of this, training should include provision for team members to try out new tools without any risk to live casework, so that they can learn what’s possible and where caution might be needed in a risk-free environment. Allowing individuals to try out the tools safely is typically much more effective at instilling understanding and building confidence (versus simply distributing training materials and policy documents, for compliance purposes).
Amgen’s approach
When Amgen looked to introduce AI within its PV operations, it applied its existing change management approach of building a “sponsorship spine” as a means to instil new behaviour, which involves modelling the desired new practice at every level of the organisation. Where a senior executive sets the vision and funds the program, it is the person responsible for reviewing a safety scientist’s cases each week who determines to what extent target individuals embrace and trust the proposed tool in the context of the work they do.
Amgen’s leaders and managers are encouraged to talk openly with their teams about the ways they use AI tools day to day in their own activities (this could be to help prepare a presentation, draft an analysis, or test an argument, whether using generic AI tools like Copilot or ChatGPT, or more specialist domain-specific tools). Crucially, managers are candid about occasions when AI’s suggestions or actions had to be overridden by their own clinical judgement — important in cultivating users’ own discernment about appropriate use of tools. Being open about fallibility has the added advantage of encouraging users to speak up about their own concerns, or ask questions, as these arise.
Understanding AI’s limitations, and when to revert
To ensure that teams adopt recommended practice from the start and stay within essential guardrails, it is also important to lay down some clear ground rules — parameters governing what AI can be used for, the data that AI tools are allowed to see, when a human needs to oversee something, and who will be accountable if something goes awry.
Understanding any boundaries and why they exist can help teams use AI more confidently and derive maximum benefits from the technology. (A lack of clarity, by contrast, can make people more cautious.) At the same time, users must understand that the technology and conditions will evolve over time, requiring ongoing revisions and adaptation to what is considered acceptable. In acknowledgement of this, the FDA and EMA have jointly published common principles for AI use across the medicines lifecycle, proposing validation of the technology that is proportional to the intended use (rather than a single, blanket standard)[2].
Monitoring people’s mood and mindset
As with training, change management is not a one-time activity, so it is important to track what teams are doing over time — as evidence of whether new behaviour has been embedded properly.
Amgen’s approach is to first gauge what people are saying (rather than tracking actual metrics, using dashboards). This includes looking for signs that team members are applying AI as an obligation, suggesting more of a compliance mindset than any genuine enthusiasm. Mentions of proactive use of an AI tool to prepare for a meeting, or comments that a slide deck came together more efficiently because of a shortcut they’d heard about, on the other hand, point to more intentional AI use and suggest a positive mindset.
Amgen does also measure AI use more formally, for evidence of the business benefits. It tracks whether a team is more productive than before, saves time, makes better and earlier decisions, for example, and whether people are working with more consistency. Relevant KPIs are agreed as part of the change program itself, to ensure that improvements are tangible and can be monitored. The reduction in time spent on repetitive tasks (in favour of judgement calls requiring a trained professional) provides an additional measure of success.
Seeing the bigger picture
Successful change management around AI use means leading with purpose; with rationale rather than what the technology can do. In drug safety, that purpose is ultimately about protecting patients, so it is paramount that PV teams understand AI’s bearing on this.
Being able to, and feel included in, the bigger picture will influence the way that users approach, embrace and use the technology, and the results that ensue.
*This article builds on themes covered in a recent life sciences industry podcast, which can be accessed in full here.
About the authors
- Bina Patel is Assistant VP of PV operations at global biotech company Amgen, based in London, UK. She has more than 30 years’ experience in PV.
- Tom Ambrosole is Senior VP and Head of Managed Services at life sciences tech company ArisGlobal. He is based in New York.
References
[1] Gartner Predicts by 2027, 50% of Enterprises Without a People-Centric AI Strategy Will Lose Their Top AI Talent, Gartner.com, 13 May 2026: https://www.gartner.com/en/newsroom/press-releases/2026-05-13-gartner-predicts-by-2027-50-percent-of-enterprises-without-a-people-centric-ai-strategy-will-lose-their-top-ai-talent
[2] EMA and FDA set common principles for AI in medicine development, EMA, January 2026: https://www.ema.europa.eu/en/news/ema-fda-set-common-principles-ai-medicine-development-0
