Agentic AI for Healthcare: Building Scalable Models for Production

The healthcare industry is undergoing a transformative shift as artificial intelligence moves from isolated predictive models to autonomous, goal-driven agents capable of reasoning, planning, and adapting in real time.

Agentic AI for Healthcare: Building Scalable Models for Production is the first comprehensive technical guide that focuses specifically on designing and deploying agentic AI systems within complex clinical and operational healthcare environments. Unlike traditional AI systems that passively infer outcomes, agentic AI involves systems that actively make decisions, coordinate with other agents, and take initiative toward long-term goals, making them ideal for dynamic and high-stakes healthcare settings. What sets this book apart is its deep technical focus tailored for computer scientists, architects, and AI engineers who are building real-world healthcare solutions. Each chapter bridges theoretical foundations with production-ready implementations using cloud-native technologies, distributed architectures, knowledge representation models, and secure deployment pipelines. With detailed case studies in oncology, vaccine management, and hospital operations, the book demonstrates how agentic AI can improve care delivery, reduce cognitive burden on clinicians, and automate routine workflows – all while meeting regulatory standards.

As healthcare increasingly adopts AI-first strategies, this book fills a critical need by offering a structured, engineering-centered approach to building scalable, intelligent, and trustworthy healthcare systems.

Sairohith Thummarakoti is a technology leader who builds agentic AI systems that work in the real world. For more than a decade, he has delivered large-scale solutions across healthcare, health insurance, banking, finance, and fintech, using Pega Cloud to orchestrate planners, trusted tools, memory with provenance, and rigorous supervision. His focus is practical: designing cloud-native architectures, streamlining enterprise workflows, and embedding intelligence into mission-critical systems while maintaining security, performance, and cost efficiency.

In the Pega ecosystem, he’s a Featured Member (top 52 among 764,000+ members) and a Top Contributor with over 500 accepted solutions spanning AI integration, decisioning, advanced design patterns, high availability, Pega Cloud 3 migrations, and secure application programming interface (API) integrations.

His agentic work centers on goal-pursuing software that perceives context, plans multi-step workflows, acts with approvals, explains decisions, and learns without drifting – always under governance. He has translated the agentic pattern – planner → tools → memory → supervision – into production services for healthcare and financial services. He treats autonomy as a dial, set deliberately to the lowest level that achieves the goal and only raised with evidence, guardrails, and human oversight. In practice, his systems have improved throughput and productivity, reduced operational cost, and tightened reliability in live applications. His deployment playbook includes equity and safety surveillance, uncertainty gating, typed tool contracts, and an incident response that is capable of pausing automation instantly.

Beyond delivery, Sairohith contributes to the global research and education community. He has authored IEEE conference papers, delivered invited talks, and presented at international forums on AI and cloud computing. As the founding Chair of the IEEE Computer Society – Columbia Section, he has organized conferences and workshops and delivered multiple Faculty Development Programs on agentic AI, helping faculty and students turn planner-tool stacks, human-in-the-loop (HITL) governance, and evaluation beyond accuracy into labs and capstone projects. He has worked as a book editor and authored several books on agentic AI, generative AI, and cloud infrastructure. He also mentors teams transitioning from prototypes to governed services. His throughline is consistent: Build systems that are explainable, auditable, and interruptible so they earn adoption where it matters most.