Artificial intelligence adoption is not a tech challenge only – it is a people journey and the organisations that master change are the ones that unlock real value.

Lisa Chamberlain James at Trilogy Writing & Consulting, an Indegene Company, and Angela Russell Winnier at Pfizer

The rapid rise of artificial intelligence (AI) in professional settings, especially in medical writing and related fields, has brought both excitement and anxiety. AI offers significant benefits such as improving workflow efficiency, enabling rapid research synthesis and making complex information more accessible. However, it also introduces new risks: subtle errors, diminished nuance and the burden of constant verification. Poorly implemented, AI can create more work, not less. As with any transformative technology, the key to successful integration lies not solely in the technology itself, but in how organisations lead and support the change, ensuring clear communication, structured engagement and robust change management throughout the process.

What do professionals want from AI and what are they getting?

The clamour for more efficient, accurate and supportive tools is growing. Surveys show that while many professionals have experimented with AI, a significant minority remain sceptical or have not tried it at all.1,2 Enthusiasm for AI is accelerating but discipline is needed to align operational reality with these high expectations. Overhyped claims can damage trust while careful deliberations and planning can align AI capabilities with real work needs. The best approach to increase trust in AI reliability and accuracy is to encourage open discussion, learning with experimentation and robust governance planning.

However, recent research shows a stark reality: while 40% of organisations have deployed AI tools, only 5% have managed to integrate them into workflows at scale.3 Most projects stall in pilot purgatory, with 95% of generative AI (GenAI) pilots delivering zero return on investment.3 This ‘GenAI divide’ highlights the challenge of moving from experimentation to real, scaled impact. Successful organisations anchor AI use cases to specific, measurable problems, not abstract innovation goals.

Why is good change management important? What does it mean for teams?

When AI is introduced, professionals often feel uncertain about their roles and the future of their work. This can lead to disengagement or resistance. Change management is about addressing these fears, providing clear information and supporting people through the transition. It’s not about replacing expertise, but about empowering professionals to use new tools effectively, keeping human judgement at the centre. Change management provides the structure that helps users understand why AI is being introduced, how it changes expectations and what success looks like in practice.

Effective AI adoption depends far more on people, process and operating model design than on technical capability. It demands a radical rethink of processes, operating models and culture. Companies that have achieved sustainable transformation have done so by simplifying strategies, redesigning processes from the ground up, and investing in capability building and leadership commitment. Change management is therefore not an accessory to AI transformation; it is the mechanism that turns capability into sustained value.

What should be available for teams and when?

Change management should be embedded throughout the AI adoption process – from early exploration and pilot projects, through to full-scale implementation and ongoing support. The goal is not simply to introduce new tools, but to build confidence, capability and trust over time. This means:

• Engaging stakeholders early with structured listening mechanisms such as polls, questions and answers, and open forums to surface concerns, understand readiness and measure sentiment. Clear communication at this stage should emphasise what AI is intended to support, what it cannot replace and how human expertise remains central
• Piloting new tools in small, low-risk contexts that allow teams to experiment safely and provide feedback before broader rollout. Feedback loops, visible iteration and transparent decision-making build trust and reduce resistance as adoption expands
• Providing role-specific training, especially for those most affected by change (eg, editors, freelancers). Training should focus less on abstract AI concepts and more on applied use in real work, reinforced through hands-on sessions, peer support and readily available guidance. Ongoing support mechanisms, such as digital champions, drop-in sessions and hyper care, help sustain confidence beyond initial launch
• Ensuring secure integration and compliance with data protection standards, reinforcing that responsible use is foundational, not an afterthought.

Leaders should also focus on building autonomous, cross-functional teams, rigorous planning, and a culture of bold innovation and collaboration. Real-time collaboration with external partners and strong vendor relationships are increasingly important for seamless execution but must be grounded in shared accountability and realistic expectations rather than assumptions of effortless automation.

The challenges in adopting AI

Realising meaningful gains from AI requires far more than deploying new technology. Adoption depends on the availability of standardised, high-quality inputs, harmonised processes and a culture of trust. Tools should reinforce the value of human expertise, not undermine it. Features such as visible tracked changes, data protection and transparency are prerequisites for confidence and sustained use.

A persistent challenge in AI adoption is the tension between automation and accountability. In regulated or high stakes environments, professionals remain responsible for decisions regardless of how much AI support is applied. This creates what is often described as a ‘verification tax’: the time and effort spent double-checking AI outputs because the system’s confidence is not well calibrated.4 When AI presents incorrect information with high confidence, the cost of error can outweigh many successful uses. This erodes trust in the promised efficiency gains.

Addressing this challenge requires more than incremental model improvement; effective AI systems must be transparent about their uncertainty, flag context gaps and learn from corrections – what some call the ‘accuracy flywheel’.4 This dynamic shifts AI from a tool that must be constantly doublechecked to one that meaningfully supports professional judgement.

Finally, adoption challenges vary by organisational context. Freelancers and small organisations face unique challenges, such as cost and access, but can also be nimbler in restructuring workflows around discrete automation. Larger organisations, by contrast, must contend with scale, governance and consistency. Across all settings, the common determinant of success is whether AI is positioned as an augmentative capability, one that depends on human understanding and accountability, rather than as a shortcut that promises effortless automation.

Tips for effective change management in AI adoption

• Start small and anchor to outcomes: break workflows into discrete tasks and try AI where the stakes are low and value can be measured
• Position AI as augmentation, not automation: use AI for tasks where it adds value and keep control where judgement matters. Avoid framing AI as a ‘single-button’ solution that bypasses professional expertise
• Invest in upskilling and sustained support: provide role-specific training aligned to real work, reinforced through hands-on practice, peer learning and accessible support. Encourage experimentation
• Be transparent and calibrate expectations: communicate openly about both the benefits and limitations of AI. Demand systems that can surface uncertainty and improve with feedback
• Celebrate credible progress, not hype: highlight incremental improvements and practical success stories that demonstrate real value
• Protect data and accountability: ensure secure integration, robust governance and compliance with data protection standards
• Modernise with intent: invest in integrated, data-centric systems and scalable automation that support redesigned workflows, rather than isolated pilots.

What is the future?

AI is here to stay, but its role will continue to evolve. The future of AI adoption is not about the wholesale replacement of professionals, nor about forcing new technology into unchanged processes. The most successful organisations will be those that empower their people to use AI as another tool in their professional toolkit: one that supports productivity and quality, but never substitutes for expertise, judgement or accountability. By focusing on transparency, incremental gains and human-centred workflows, organisations can harness AI’s benefits while safeguarding expertise and quality. As organisations move beyond early experimentation, the focus will shift from novelty to reliability, from what AI could do to what it demonstrably does well. This means embracing AI systems that are transparent about their limitations, calibrated in their confidence and designed to improve through sustained human feedback. Trust, once earned, becomes a durable enabler of scale.

Industry leaders are already showing that when AI adoption is grounded in process redesign, continuous learning and integration into real work, meaningful value follows. The lesson is clear: demand a different kind of AI, one that is tightly integrated into workflows, transparent about its limits and capable of learning with every interaction. This disciplined approach, rather than hype-driven adoption, is what distinguishes the small proportion of organisations that achieve lasting impact from those that stall after early enthusiasm.5

Medical writers are uniquely positioned to lead these initiatives, ensuring that AI adoption is thoughtful, measured and truly fit for purpose.

References:

1. Visit: salesforce.com/news/stories/customer-engagement-research2023/?utm_campaign=WhyWereNotAddingAItoPerfectIt&utm_content=Jan24&utm_source=Blog
2. Visit: amwajournal.org/index.php/amwa/article/view/528
3. Visit: forbes.com/sites/jaimecatmull/2025/08/22/mit-says-95of-enterprise-ai-failsheres-what-the-5are-doing-right/
4. Visit: forbes.com/sites/jasonsnyder/2025/08/26/mit-finds95-of-genai-pilots-fail-becausecompanies-avoid-friction/
5. Visit: hbr.org/2025/11/overcoming the-organizational-barriers-to-ai adoption

Angela Winnier joined Pfizer in 2018 and is executive director of Medical Writing and the therapeutic area lead for internal medicine and inflammation/immunology. In addition to her pipeline support, Angela serves as technology lead for medical writing, providing strategic and operational oversight to achieve successful technical implementation, change management and operational efficiencies. Angela has spearheaded key technology and AI‑assisted authoring initiatives and has played a critical role in upskilling Pfizer research and development teams.

Dr Lisa Chamberlain James is a senior partner and VP Medical Solutions at Trilogy Writing & Consulting, an Indegene Company. Aside from management activities, she leads client projects, with extensive experience in a variety of documents. Lisa has a special interest in writing for the public, and in the effective and ethical use of AI in medical writing. Following a PhD and postdoctorate in Pathology at Cambridge, UK, Lisa began her medical writing career in 2000. Since then, she has been involved in the European Medical Writers Association (EMWA) as a member of the Educational Committee, mentor, leader and assessor of workshops, and is a member of the American Medical Writers Association (AMWA)’s Executive Forum, AI Task Force, and teaches and reviews workshops for AMWA. Lisa regularly takes part in podcasts, articles and is a member of several global, pan‑industry panels and committees, holds an EMWA professional development certificate, is a visiting lecturer for King’s College London, UK, and the University of Lisbon, Portugal, and is the current chair of the Communicating with the Public SIG, and the Geoff Hall Scholarship Committee. She is a section editor of Medical Writing, an EMWA Nick Thompson Fellow and a Fellow of the Royal Society of Medicine.