Video summary
Pharma's AI Moment: More Innovation in 12 Months Than the Previous 10 Years
Main summary
Key takeaways
Summary
Mahi Rayasam, Eli Lilly’s SVP and Information Officer, described how the company applies technology and AI across drug development and commercialization. The overarching goal is to get safe, effective medicines to patients faster and improve access.
Strategy and operating model
Rayasam leads technology for key commercial platforms, including Lilly USA and LillyDirect, as well as clinical development. This work spans the path from molecule development through trials, regulatory submissions, and manufacturing.
Lilly’s AI efforts cover discovery, clinical development, manufacturing, and commercialization. The guiding outcome is patient impact—not AI adoption for its own sake. Rayasam said Lilly applies principles similar to its asset-portfolio management approach when prioritizing AI investments.
AI prioritization playbook
- Start with the outcome: Favor opportunities that can help medicines reach patients more quickly, safely, and accessibly.
- Check whether AI is the right tool: Some processes may not benefit from AI.
- Reconsider the workflow itself: Don’t simply automate an outdated or unnecessary process; redesign the workflow where appropriate.
- Prioritize by expected impact and return: The key investment question is which use cases best advance the goal of getting medicines to patients sooner. Lilly did not disclose a specific ROI formula or threshold.
Implementation and adoption
- Keep expert judgment central: Rayasam characterizes AI as an augmentation and acceleration tool, not a substitute for scientists, engineers, or technicians. Experts remain responsible for directing work and evaluating outputs.
- Build trust through testing: Lilly does not expect scientists to trust AI automatically. Users should test systems, probe their limitations, and judge where they fit into their work.
- Enable experimentation: Technology teams provide platforms, access, and examples showing how AI might make work faster, more creative, or less constrained. Scientists can adapt those ideas to their own workflows.
- Use AI to free time for higher-value work: Examples include supporting scientific discovery, software development, product development, and the design of consumer-facing materials and websites.
Products, access, and customer experience
LillyDirect is intended to make access to Lilly medicines and related resources easier. Rayasam described it as a combination of access, education, content, guidance, and resources—not simply a way to bypass prior authorization.
The approach includes mapping the patient journey and addressing pain points. LillyDirect is intended to complement physician advice, not replace it.
Talent and leadership
Rayasam argued that innovation often comes from the intersection of disciplines, rather than from one person possessing every skill. Lilly’s teams bring together capabilities such as biology, chemistry, AI, data science, engineering, operations, cybersecurity, and pharmaceutical or medical-device experience.
Leaders’ role is to assemble complementary talent, help teams work across disciplines, and enable innovation at those intersections. Rayasam said he hires for both expertise and the ability to execute at the pace required.
Metrics and timelines
- Rayasam said he had seen more innovation in the previous 12 months than in the prior 10 years. This was a qualitative observation, not a reported operating KPI.
- He had been in his Lilly role for about one year at the time of the interview.
- Lilly was celebrating its 150th anniversary.
- No specific AI investment amounts, revenue targets, productivity gains, or adoption KPIs were provided.
- Rayasam estimated that his prior McKinsey client work reached more than 500 million patients through the healthcare organizations served.
Presenters and sources
- Brad Hairston — Host, Transform Now
- Mahi Rayasam — SVP and Information Officer, Eli Lilly and Company; interview guest and source for the Lilly strategy and practices described above.
Rate this summary
Your feedback will help improve summaries.
Improve this summary
Reprocess with a stronger model when the summary feels incomplete or inaccurate.
Translate summary in another language
Ask questions to this video
Chat for follow-up questions, clarifications, and source-backed answers.