Generative AI and agentic workflows are rapidly transforming commercial pharmaceutical operations, enabling greater agility, improved decision-making, and enhanced customer engagement. In a panel discussion moderated by Chris Walker, Director of Product Marketing at Tellius, industry leaders from Novartis and Bayer shared how these technologies are creating tangible value across multiple functions.
Sri Rao, who heads data and analytics solutions at Bayer Pharma, and Ravi Shankar, who leads data and analytics enablement at Novartis, joined Walker to discuss the current applications and future potential of AI in pharmaceutical commercial operations.
Current Applications of Generative AI in Pharma
Rao highlighted multiple areas where Bayer is currently leveraging generative AI, including field team support, brand insights, and market access. "We're helping area general managers work with their reps in terms of coaching them, educating them on where they need to focus and key talking points when discussing with doctors," he explained. "Using generative AI to generate personalized notes for coaching conversations with reps is something we're working towards."
Shankar categorized AI applications into three areas: operational efficiency, orchestrated experience, and innovative solutions. For orchestrated experiences, he described how AI can optimize field force effectiveness: "We can give an optimized call plan which can be orchestrated not only for that particular sales rep but around the targets. Whether it's an MSL, a TLL, or an FRM, it can give you the integrated plan for targeting a particular doctor."
Walker added that from Tellius' perspective, AI-powered tools should function like "an Iron Man suit" for analysts and business users, enabling them to answer not just what happened, but why it happened, identifying root causes, key drivers, and anomalies.
Behavioral Shifts and Practical Applications
The implementation of AI is driving significant behavioral changes in how pharmaceutical professionals work. Shankar described an orchestrated solution for sales representatives that plans their day, coordinates messaging across functions, and captures insights in real-time. "This whole orchestrated solution will give a cohesive experience, bring effectiveness, and bring operational efficiency," he noted.
Bayer has undergone an organizational transformation with a concept called "dynamic shared ownership," where area general managers own their micro-business. Rao explained how AI supports this model: "We worked with a team utilizing capabilities like Tellius and some Gen AI capabilities to enable them to answer questions much faster in simple plain English. They can ask follow-up questions about market share or competitors, dig deeper into key drivers, and understand the reasoning behind trends."
Technology Infrastructure Requirements
All three panelists emphasized that technology should serve business needs rather than drive them.
"Before we go into technology, we need to understand the business problem. Technology is an enabler. End of the day, the business solutions prevail, the business challenges prevail."
- Ravi Shankar, Executive Director, Data & Analytics, Novartis
Rao broke down the technological components into several categories, including foundation models like BERT or GPT-4, natural language query (NLQ) capabilities, self-service tools for analysts and field representatives, and unstructured data analysis. "You don't need to be too technical to analyze stuff. You can ask a question about profitability or understanding the performance of a particular brand versus your competitor in simple natural language," he explained.
Walker highlighted Tellius' approach to providing natural language interfaces to enterprise data, noting that "the key differentiator is that unlike a generic Gen AI platform, Tellius has put the time into making that semantic and ontology layer and connecting it to our customer's situation so that the answers come back in a nuanced, contextual way."
The Future of Commercial Pharma with AI
Looking ahead, the panelists see tremendous potential for AI agents to transform commercial operations. Rao envisioned a "rep buddy" that could follow representatives during doctor visits, understand conversations, and provide feedback for follow-up interactions. He also described how agents could mimic skilled analysts: "You could have agents built that can combine when you ask a question through a simple NLQ-based question. The master agent should be able to go and get details and give you the right answer without waiting for an analyst."
Shankar emphasized how AI will make organizations more agile and nimble, particularly in connecting disparate data sources to create comprehensive patient journeys. "If I can use EMR/EHR data, which is semi-structured and sometimes in PDF format, and join that data with our traditional data sources like claims, we can have an end-to-end patient journey created," he explained.
Walker observed that the themes emerging across the discussion centered on "patient outcomes and commercial efficacy," noting the wide range of use cases where AI could make a significant impact.
Overcoming Implementation Challenges
Survey results from the audience revealed that the biggest challenges to AI implementation are lack of technical know-how or bandwidth, followed by unclear business outcomes. Shankar acknowledged the rapidly evolving nature of AI models and recommended a model-agnostic approach: "We need to have a model-agnostic approach and start learning the basics with expertise from the industry who can help us understand and get us the bandwidth."
Rao noted that while proof-of-concept phases are relatively easy, scaling to enterprise solutions presents challenges.
"Having a clear set of use cases in terms of what you're trying to solve for, not boiling the ocean, ensuring you have the right data and the right SMEs who can help you, and finding a champion from the user perspective is equally important."
- Sri Rao, Head of Data and Analytics solutions at Bayer.
Walker added that having a solid partner who's established in the AI space can provide the helping hands needed to navigate these challenges successfully.
Five Key Takeaways:
1) Focus on specific business problems: "Having a clear set of use cases in terms of what you're trying to solve for, not boiling the ocean, ensuring you have the right data and the right SMEs who can help you," advises Sri Rao from Bayer.
2) AI agents enable orchestrated customer experiences: "We can give an optimized call plan which can be orchestrated not only for that particular sales rep but around the targets. Whether it's an MSL, a TLL, or an FRM, it can give you the integrated plan," explains Ravi Shankar from Novartis.
3) Natural language interfaces democratize data analysis: "You don't need to be too technical to analyze stuff. You can ask a question about profitability or understanding the performance of a particular brand versus your competitor in simple natural language," says Rao.
4) Connected data creates comprehensive patient journeys: "If I can use EMR/EHR data and join that data with our traditional data sources like claims, we can have an end-to-end patient journey created," notes Shankar.
5) Finding user champions is crucial for adoption: "Champions become your ambassadors to help further navigate through some of these challenges," emphasizes Rao regarding successful AI implementation.
For more market access & RWE insights, including latest articles, interviews and more – click here
