Ramji Vasudevan, Business Unit Head - Life Sciences, Altimetrik

The pharmaceutical industry has moved well beyond asking whether artificial intelligence (AI) can transform drug development. According to a survey by the Pistoia Alliance,¹ 68% of life sciences R&D professionals are using AI and machine learning in their work. Yet the same survey reveals that 52% cite poor data quality as the biggest barrier to AI implementation.

This paradox captures exactly where the pharma industry finds itself in 2025. Companies have moved from wondering “what can AI do?” to conducting meaningful experiments. But most remain stuck between proof-of-concept success and enterprise-scale value.

That hurdle is not a technology problem. AI models have improved dramatically, with new capabilities emerging every few months as access becomes increasingly democratized. The potential of AI has always been there, and that fundamental promise hasn’t changed.

But McKinsey research says generative AI models account for only about 15% of a typical project effort.² The remaining 85% involves adapting models to a company’s internal knowledge base and use cases—and this is where most drugmakers struggle.

The real barriers are about business transformation. Consider a common scenario in pharmaceutical manufacturing: Drug test results are still shared via email attachments and PDFs between contract manufacturers and testing laboratories. These documents often contain handwritten notes that get scanned into PDFs. Then someone has to manually type out the handwritten notes so they can enter them into their own system. It’s slow and error-prone.

Working with numerous pharmaceutical companies on AI implementation has revealed a clear framework for moving from successful experiments to enterprise-scale value. The approach centers on business outcomes rather than technology capabilities.

Step 1: Start with clear business outcomes

Too many organizations begin with the technology and then search for applications. Business outcomes must drive the digitization and use of tools such as AI. A single IT function can’t effectively solve problems across all the different business domains.

Instead, start by identifying specific, measurable business problems. Consider clinical study report generation: from the last patient’s final visit to the clinical study report generation, it takes anywhere from a few weeks to multiple months. AI can handle 60%-70% of document creation much faster than humans, while still maintaining human oversight. Don’t shoot for 90%-95% at the beginning. Begin with modest automation goals and iterate toward optimization.

Step 2: Create joint accountability

One of the biggest barriers to AI scaling is the traditional divide between business and IT functions. Siloed approaches consistently fail to deliver enterprise value. According to McKinsey, 70% of transformations fail, in part because of the disconnect between what the business wants and what the IT teams deliver.³

Gartner recently suggested putting the chief information officers (CIOs) and the chief experience officers (CXOs) of the business functions together and holding them jointly accountable for outcomes.⁴ Each leader owns their domain, but shares ownership of results. This creates true partnership rather than traditional handoffs. Gartner says that nearly three-quarters of CXOs who co-lead delivery meet or exceed targets from digital investments.

Step 3: Implement hub-and-spoke governance

Successful AI scaling requires balancing centralized governance with federated execution. Central functions should maintain oversight of safety, regulatory compliance, and enterprise-wide standards—particularly crucial in the heavily regulated pharmaceutical sector.

However, use case development should be federated to business units. Some responsibility should be co-owned between the CIO and CXO, while safety and regulatory aspects remain centralized at the enterprise level. This ensures that business units can move quickly while maintaining necessary controls. Eventually, companies should work toward democratized AI access where business users can build their own solutions, test them, and either scale them to production or discard them based on results.

Step 4: Build modular solutions

Rather than attempting wholesale digital transformation, focus on bolt-on solutions that don’t require overhauling entire systems. The document processing example, for instance, doesn’t need any fundamentally new system. It can be a bolt-on with possibilities of what AI brings to the table.

Merck and McKinsey co-developed a platform that generates clinical study reports.It reduced the time to produce a first draft from 180 hours to 80 hours while cutting errors in half.⁵ It’s a clear opportunity for AI augmentation without requiring complete infrastructure replacement.

Step 5: Plan for democratization

The ultimate goal is creating infrastructure in which business users can access data, compute power, and AI models directly. With that infrastructure comes a drastic change in how organizations operate. Anyone with the right credentials should be able to access data, compute, storage, and AI models, connect analytics, and make informed business decisions in real-time. The impact is almost instantaneous. Instead of three-month wait times for simple dashboards, business users can create solutions over an afternoon. But this democratization must be built on solid data foundations—achieving a true single source of truth rather than maintaining data silos.

The competitive advantage ahead

While many organizations remain focused on experimenting with models and use cases—the 15% of the effort—the real value lies in solving the 85% challenge of business adaptation and integration. Companies that master this business-first approach to AI scaling will create sustainable competitive advantages.

The goal is enabling organizations to operate at the speed of business decisions rather than the speed of IT implementation. Doing so can bring life-saving treatments to patients faster, moving from months-long processes to real-time insights and actions.

Success requires moving beyond the technology-first thinking that has characterized much of the industry’s AI journey. The companies willing to do the hard work of business transformation will be the ones that capture that potential at enterprise scale.

Ramji Vasudevan is Business Unit Head - Life Sciences, Altimetrik

References

1. What will the Labs of 2030 Look Like? Pistoia Alliance. https://marketing.pistoiaalliance.org/hubfs/Lab%20Of%20The%20Future%20Reports/Lab%20Of%20The%20Future%20Survey%20Results%202024%20.pdf

2. Generative AI in the Pharmaceutical Industry: Moving From Hype to Reality. McKinsey & Company. January 9, 2024. https://www.mckinsey.com/industries/life-sciences/our-insights/generative-ai-in-the-pharmaceutical-industry-moving-from-hype-to-reality

3. Perspectives on Transformation. McKinsey & Company. https://www.mckinsey.com/capabilities/transformation/our-insights/perspectives-on-transformation

4. Fostering CIO Partnerships With CxOs Takes on New Urgency. Gartner. https://www.gartner.com/en/chief-information-officer/insights/cio-partnerships
5. With Gen AI, Merck and McKinsey Transform Clinical Authoring. McKinsey & Company. https://www.mckinsey.com/about-us/new-at-mckinsey-blog/with-gen-ai-merck-and-mckinsey-transform-clinical-authoring