Peer Review: AI in Biopharma – From Vision to Impact
In 2025, the biopharma industry has turned a corner from experimentation with AI to real implementation to address critical biopharma needs. In a recent Peer Reviews session, Anita Modi, CEO and co-founder of Peer AI , sat down with Bharathram Lakshmivarahan , Partner at Slalom in the Life Sciences Practice, to explore how biotech and pharma leaders are driving AI adoption and overcoming barriers to deliver real ROI.
Key sound bytes from the conversation:
🔍 What’s different about AI in 2025? Bharathram shared that 2024 was about pilots, experimentation, and getting familiar with AI. In contrast, 2025 is a “prove it” year. We're seeing a lot of executives starting to mandate to the business an understanding of what the value of AI is really going to be. Companies are being asked to quantify value—if a use case can’t show ROI within 12–24 months, it’s unlikely to get funded.
We’re also shifting from summation and generative AI to agentic AI—systems that can make autonomous decisions.
🚀 Most promising AI use cases in pharma today?
Bharathram focused on a few cases across drug development:
R&D Impact: Patient recruitment is being revolutionized. AI helps find “needles in a haystack” by matching the right patients to the right trials—essential in an era of personalized medicine.
Supply Chain Transformation: COVID exposed major weaknesses here. Now, AI is helping with predictive inventory management and demand forecasting.
Elevation of Medical Affairs: Medical Affairs is finally elevated to that third pillar in biopharma, alongside R&D and commercial. AI is streamlining data synthesis—from CRMs, real-world evidence, and clinical data—into literature reviews, disease state summaries, and HCP materials.
💡 The ROI Conversation: Not your Typical Spreadsheet
Bharathram framed two main schools of thought:
Traditional ROI calculators—analyzing implementation costs vs. expected gains.
ROI calculator is dead—less emphasis on financial projections, and a more nuanced approach focused on strategic business value and outcomes.
There’s no right answer here. Some leaders are saying, if we don’t see X amount of value in 12 to 24 months, we’re cutting the project. Regardless of approach, organizations need a standardized, enterprise-wide framework. "It’s not commercial having its fiefdom and medical affairs having secret sauce. We need a uniform methodology across the company."
Bharathram and Anita also emphasized that ROI doesn’t always come from a big bang implementation. "It’s a lot of incremental support that builds the value story."
🧩 What kind of operating environment supports successful AI implementation?
Based on what Bharathram sees day to day, the key is the operating model.
One one end, we see a Centralized Model – where A single leadership entity sets enterprise-wide AI governance standards, ensuring consistent application but may limit rapid response to new trends.
The other end of the spectrum is a completely decentralized model – Autonomous business units independently govern AI, fostering innovation and responsiveness, albeit at the risk of siloed approaches and potential misalignment.
We’re likely moving toward a more Federated model – Merging overarching corporate policies with departmental flexibility, this approach centralizes control while decentralizing for scalable innovation and agility.
Governance is key. Bharathram recommends including compliance, legal, IT, and business stakeholders from day one.
Finally, Bharathram highlighted: "It’s very important to have a team that has a pulse on the regulatory environment."
🤖 Build or Buy? The Eternal Dilemma?
Bharathram’s advice? Ask one question: “Do we have the data, capability, and time to build something better than what’s available in the market?” And it’s an AND, not an OR—all three need to be true.
He emphasized the complexity of even defining "buy": it could mean deployment tools, AI-as-a-service models, or plug-ins. "It’s free to go look. The industry is moving fast, and the landscape changes every quarter."
Interesting Trend Note: Anita shared that Peer AI has seen the “build boomerang” – companies who opted to build last year, but are coming back to buy given the complexity. "They hit real challenges staying up to date with the newest technology or getting enough value, fast enough."
Both agreed: even if you don’t move forward with a vendor now, starting the conversation helps you learn.
🧭 Leadership Advice for Navigating AI
Bharathram kept it simple:
Be a student. Stay curious—it’s moving too fast to stay still.
Talk to the business. Learn their pain points and blockers—tech, process, data.
Avoid ego. The best ideas come from open, cross-functional dialogue.
🛡️ Addressing the Elephant in the Room: Job Security
Bharathram sees this a lot on the ground. His advice? Invest in AI literacy across your org. Then do a task-level analysis: What should stay human vs. what can be machine-assisted? AI should be seen as a force multiplier, not a threat.
The 2025 Imperative
This is the year to move from AI experimentation to strategic implementation. Your competitive advantage lies in:
Outcome-focused approaches
Robust foundational strategies
Continuous learning
🚀 Remember: In the fast-evolving world of AI, your strategy today might look completely different next quarter. Start somewhere, start learning, start now.
Insights from a conversation between Anita Modi, CEO of Peer AI , and Bharathram Lakshmivarahan, Partner at Slalom Life Sciences Practice.