Assistive vs Agentic AI in Life Sciences: Use Cases & Learnings

In the latest episode of The Intelligent Leader Podcast, host Shashank Garg sits down with Kabir Patel, Global Head of IT at AstraZeneca, to unpack the evolving landscape of AI in enterprise settings—specifically, the growing shift from assistive to agentic AI. With a background spanning technology leadership at Oracle, Thomson Reuters, and now global pharma, Kabir brings an insightful perspective on how AI is driving digital transformation across life sciences.

Kabir and Shashank dive deep into the definition of agentic AI—what it is, what isn’t, and why it matters now more than ever. Unlike assistive AI or traditional automation, agentic AI is designed to reason, act, and learn independently, pushing past human-in-the-loop models to enable autonomous systems. Kabir emphasizes that not all AI that claim to be “agentic” meets the bar. True agentic systems don’t just respond to commands—they interpret context, deconstruct complex tasks, and execute across multiple domains with minimal human input.

One standout insight from the episode is the potential of agentic AI to build “self-healing” systems, particularly in pharma’s complex supply chains. Imagine AI agents that detect failures, simulate corrective actions, learn from past disruptions, and autonomously implement solutions—radically reducing downtime and operational cost. Kabir argues that these use cases aren’t new; what’s changing is the level of autonomy and adaptability now possible, thanks to recent advancements in AI architectures and infrastructure.

Yet, with this promise comes the reality of readiness. Kabir and Shashank stress that organizations shouldn’t jump straight into building agentic systems if they haven’t yet experimented with assistive AI. The maturity curve matters. Teams must start with pilot projects in low-risk functions like procurement, HR, or IT, using them as testbeds for foundational AI capabilities such as memory, context awareness, chaining, and integration across tools and platforms.

Scaling agentic AI also requires more than just good technology. It demands cultural buy-in, trust in AI-driven decisions, and an enterprise-wide willingness to reimagine how work gets done. Kabir outlines the core ingredients for success: quality data, strong use case prioritization, infrastructure for AI operations (AIOps), and most importantly, a mindset shift from standardized processes to personalized, adaptive workflows.

Looking ahead, the conversation touches on how the enterprise tech stack is evolving. With nearly every major vendor—from cloud giants to SaaS providers—launching agentic AI capabilities, companies face critical build-vs-buy decisions. Kabir predicts that most organizations will adopt a hybrid approach: leveraging off-the-shelf platforms for speed, integrating with existing systems for efficiency, and custom-building where unique value lies.

Beyond enterprise software, Kabir also imagines a future where agentic AI drives personalized experiences at scale—where every user interface, decision system, and digital workflow adapts to the individual. In this world, personalization isn’t a luxury—it’s the new baseline for productivity and innovation.

Agentic AI is not a distant vision—it’s already reshaping how life sciences and global enterprises operate. For leaders ready to move beyond pilots and proofs of concept, this episode offers a rich framework for understanding where to begin, what to build, and how to lead.

Want to dive deeper into this conversation? Listen to the full episode of The Intelligent Leader Podcast to explore how AstraZeneca and other organizations are building the next generation of intelligent systems.