Artificial intelligence has moved well beyond the experimental phase in biopharma, according to Chris Meier, managing director and partner at Boston Consulting Group, in a presentation made at the EHIA’s Investors in Healthcare Private Capital Conference in June.
While much of the public discussion continues to focus on the capabilities of large language models and generative AI, leading pharmaceutical companies are already deploying AI across research, manufacturing, commercial operations and corporate functions to deliver measurable improvements in productivity, cost efficiency and revenue growth.
The next challenge is not whether AI works, but how organisations scale it. Companies that treat AI as another technology project are unlikely to realise its full value but those that redesign processes, invest in data and people, and embed AI into their operating model are beginning to create sustainable competitive advantage.
For investors, boards and executives, the question is no longer whether AI will transform biopharma, but how quickly companies can move from isolated applications to AI-first ways of working.
Value creation
The first wave of AI adoption has produced tangible results across the pharmaceutical value chain: AI-assisted molecular design is accelerating discovery and improving the efficiency of clinical trial execution; commercial teams are using agentic AI to increase sales force productivity and deliver more personalised engagement with healthcare professionals; manufacturing organisations are applying machine learning to optimise production processes, increasing yields and reducing variability; and corporate functions are deploying AI to improve forecasting, automate routine activities and enhance decision-making.
The impact is increasingly quantifiable: R&D – 20–45% faster molecule design and 10–20% faster clinical trial execution; commercial – 20–30% improvements in field-force efficiency; operations and supply chain – 15–30% improvements in manufacturing output; and corporate functions – 40–50% improvements in financial forecasting accuracy.
Collectively, they demonstrate that AI is no longer an experimental technology but a practical business capability that can improve both productivity and financial performance.
The next stage: transformation not automation
Many organisations begin their AI journey by automating existing tasks. While valuable, this represents only the first stage of maturity.
The presentation introduced a useful framework describing three progressively more ambitious approaches:
Deploy: Focused on improving existing workflows through AI-enabled automation through drafting regulatory submissions, generating medical content, summarising documents or automating repetitive administrative tasks.
These initiatives improve productivity but largely leave underlying business processes unchanged.
Reshape: Moving beyond task automation to redesign end-to-end workflows. Rather than simply accelerating individual activities, AI is embedded within operational processes, such as with biologics manufacturing, where AI analyses historical fermentation data to identify the drivers of high-yield production batches. Initial recommendations improved manufacturing consistency, before real-time sensors and predictive models were introduced to optimise process settings continuously. The result was a 20–25% improvement in production yield, generating tens of millions of dollars in additional value.
The same philosophy is emerging across commercial operations, where AI can coordinate physician engagement, personalise communications and coach sales representatives before customer interactions.
Invent: The greatest long-term opportunity lies in creating entirely new operating models.
Drug discovery illustrates that shift well. Traditional research relies on laboratory experiments to generate scientific knowledge before designing candidate molecules. An AI-first “dry lab” reverses this relationship.
AI systems generate and optimise thousands of candidate molecules computationally, while robotic laboratories conduct targeted experiments primarily to generate new data that improves the models. Rather than scientists driving discovery supported by AI, AI increasingly drives discovery supported by experimental validation.
Although still emerging, this approach illustrates how AI may fundamentally redefine pharmaceutical R&D over the coming decade.
The opportunity beyond productivity
AI’s value extends far beyond cost reduction.
The presentation estimated meaningful opportunities across both revenue growth and operating efficiency.
Revenue opportunities include: accelerating drug discovery; improving clinical trial design and patient recruitment; delivering personalised commercial engagement; optimising pricing and contracting; and improving manufacturing planning.
Cost reductions arise through: more efficient clinical trial execution; automated medical and regulatory content generation; predictive maintenance; inventory optimisation; contract review and corporate automation.
The largest revenue opportunities are concentrated within R&D and commercial operations, while manufacturing and corporate functions provide substantial efficiency gains.
For investors, AI should increasingly be viewed not simply as a cost-saving technology but as a driver of future competitive advantage and revenue growth.
AI leaders
Not every organisation is progressing at the same pace, but leading AI adopters share several common characteristics.
Rather than pursuing hundreds of disconnected pilot projects, they focus on a small number of strategically important initiatives. They invest heavily in data infrastructure rather than viewing AI models as standalone solutions, they build specialised AI teams while simultaneously upskilling the wider organisation, and most importantly, they increasingly approach business challenges with an AI-first mindset – asking how AI could redesign a process before assuming existing workflows should remain unchanged.
The biggest differentiator between companies is not geography – some European companies are leading adoption while some US companies are lagging – instead, organisational willingness to embrace risk appears to be a stronger predictor of progress.
Technology only part of success
Companies often become preoccupied with selecting AI models or experimenting with the latest foundation models. Research suggests this emphasis is misplaced.
Successful AI transformation depends roughly on three elements following a 10/20/70 rule: 10% selecting and configuring the right algorithms.; 20% building robust technology platforms and high-quality data infrastructure, and 70% redesigning organisational processes, operating models and ways of working.
The implication is that competitive advantage will come not from access to AI models, which are increasingly available to everyone, but from how effectively organisations redesign themselves to use those models.
It was an observation reinforced repeatedly.
Successful deployment often depends less on the sophistication of the model than on the quality of data, systems integration and user adoption.
Lessons from the discussion
The general discussion offered several practical insights into how AI adoption is evolving.
Regulatory acceptance increasing – Around 20–30% of pharmaceutical companies have already submitted regulatory dossiers that were drafted using AI before undergoing human review. Regulators have generally accepted this approach, provided human accountability remains in place.
Interestingly, regulators are also adopting AI themselves. Agencies are beginning to use AI to review lengthy submissions, raising the longer-term possibility that future regulatory processes could become increasingly data-centric rather than document-centric.
Data remains biggest technical constraint – Organisations frequently underestimate the effort required to curate historical datasets, establish real-time data feeds and integrate AI tools into existing systems.
Organisational resistance remains significant – Technology can often be deployed relatively quickly, but achieving widespread adoption across an organisation remains more difficult.
Talent strategies evolving – Many companies have found greater success teaching AI to experienced pharmaceutical professionals than attempting to teach pharmaceutical science to AI specialists.
Infrastructure concentration presents emerging risks – As reliance on a small number of cloud and AI providers increases, organisations are becoming more conscious of platform dependency, performance variability and resilience.
These operational realities may prove more important than headline announcements about new AI capabilities.
Five priorities for leadership
The presentation concluded with five practical recommendations for organisations seeking to scale AI successfully.
- Leaders should move beyond incremental automation and focus on redesigning entire processes around AI and autonomous systems.
- Organisations should commit meaningful capital to a limited number of strategic AI initiatives rather than dispersing resources across numerous pilots.
- AI investments should be evaluated against measurable business outcomes – including trial speed, manufacturing yield, commercial performance and patient benefit.
- Data should be treated as strategic infrastructure requiring governance, ownership and sustained investment.
- AI success depends upon attracting scarce talent while establishing clear executive accountability and governance.
Conclusion
The biopharma industry is approaching an inflection point.
The first generation of AI projects demonstrated that productivity gains were possible. The next generation is beginning to redesign workflows, reshape operating models and redefine how medicines are discovered, developed and commercialised.
The message is clear, competitive advantage will belong not to the organisations experimenting with the largest number of AI pilots, but to those prepared to rethink how their businesses fundamentally operate.
In the coming years, AI is likely to become less a standalone technology initiative and more the operating system underpinning the entire pharmaceutical enterprise.
