Agentic AI is rapidly emerging as one of the most transformative technologies in healthcare.

 

Before joining our AI in Healthcare Conference, take a few minutes to read this exclusive interview with Sanjib Basak, CTO, SaffronHealth.AI, who will be speaking on: "Agentic AI in Healthcare: Eliminating $1T in Waste. Lowering Costs. Improving Patient Outcomes"

 

1. Can you tell us about your background and what led you to focus on Agentic AI in healthcare?

Ans: My background spans more than two decades in healthcare analytics, AI, and population health, where I've had the opportunity to lead teams and large-scale transformation initiatives at IBM Watson Health, Optum, and Alight Solutions.

Throughout my career, I noticed a consistent challenge: healthcare generates enormous amounts of data—claims, EHRs, pharmacy data, lab results, and social determinants of health—but very little of that data is translated into actionable recommendations. Most analytics platforms stop at generating reports, dashboards, or risk scores, leaving clinicians, care managers, and administrators to determine what to do next.

At the same time, healthcare costs continue to rise at an unsustainable rate, yet consumers often receive very little guidance in navigating their healthcare journey. Patients are expected to make complex decisions about providers, treatments, medications, benefits, and care plans with limited support. The result is delayed care, missed preventive opportunities, poor adherence, unnecessary utilization, and ultimately worse health outcomes. Despite spending more than any other country on healthcare, we often fail to help individuals make the right decisions at the right time.

That realization led me to focus on Agentic AI. I believe healthcare doesn't have a data problem—it has an execution problem at scale. We already know many patients who are at risk, which care gaps exist, and where waste occurs. The challenge is coordinating millions of decisions and interventions across patients, providers, care managers, and health systems.

Agentic AI introduces a new paradigm where specialized AI agents can continuously analyze data, identify opportunities, coordinate actions, and support decision-making in real time. Instead of simply telling us who is high risk, these agents can help determine why they are at risk, what intervention is most likely to work, who should engage the patient, and even guide the consumer toward the next best action.

What excites me most is the potential impact. The U.S. healthcare system spends trillions of dollars annually, with estimates suggesting that more than $1 trillion is lost to administrative inefficiencies, fragmented care, preventable utilization, and delayed interventions. While much of today's AI is focused on improving productivity and automating existing processes, I believe Agentic AI gives us an opportunity to solve some of healthcare's most fundamental challenges—improving outcomes, lowering costs, empowering consumers, and making healthcare more proactive rather than reactive.

That's what ultimately drew me to this space and inspired me to build solutions that move beyond prediction and reporting toward autonomous, coordinated action. The future of healthcare isn't just better insights—it's intelligent systems that help every stakeholder—from patients to providers to payers—make better decisions and take action at the right moment.

 

2. How is Agentic AI different from traditional automation, chatbots, or AI copilots currently being used in healthcare?

Ans: That's a great question because there's a lot of confusion in the market today around automation, copilots, and Agentic AI.

Traditional automation is rules-based. It follows predefined workflows and can only handle scenarios that have been explicitly programmed. It's very effective for repetitive tasks but struggles when situations become complex or require judgment.

Chatbots represent the next evolution. They can interact with users in natural language and answer questions, but they are largely reactive. They wait for someone to ask a question and then provide a response. They don't typically understand broader context, initiate actions, or coordinate across systems.

AI copilots go a step further by assisting humans with specific tasks. For example, a physician copilot might summarize a patient chart, draft clinical notes, or suggest potential diagnoses. These tools improve productivity, but the human remains responsible for deciding what action to take and coordinating the workflow.

Agentic AI is fundamentally different because it is goal-oriented rather than task-oriented. Instead of simply answering questions or assisting with a single activity, a network of specialized AI agents can work together to achieve an outcome.

For example, imagine a diabetic patient who is at high risk for hospitalization. A traditional analytics platform might generate a risk score. A copilot might summarize the patient's history for a care manager. An Agentic AI system, however, could identify the risk, determine the key drivers behind that risk, detect care gaps, evaluate available interventions, recommend the next best action, prioritize outreach, coordinate with care management workflows, monitor outcomes, and continuously adapt its recommendations as new information becomes available.

In healthcare, this distinction is incredibly important because most of our challenges are not information problems—they are coordination problems. Patients interact with providers, specialists, pharmacies, health plans, employers, and care managers, all operating in separate systems. Agentic AI has the potential to act as an intelligent orchestration layer that continuously coordinates decisions and actions across those stakeholders.

The way I often describe it is:

• Automation executes tasks.

• Chatbots answer questions.

• Copilots assist people.

• Agentic AI pursues outcomes.

That's why I believe Agentic AI represents a much bigger opportunity for healthcare. It moves us from simply generating insights or improving productivity to actually helping healthcare organizations achieve measurable outcomes such as reducing hospitalizations, closing care gaps, improving member experience, and lowering the total cost of care.

 

3. What are some real-world use cases where Agentic AI can reduce costs while simultaneously improving patient outcomes and care quality?

 

Ans: One of the reasons I'm so excited about Agentic AI is that it creates a rare opportunity in healthcare: reducing costs while improving outcomes and quality at the same time. Traditionally, organizations have viewed those goals as competing priorities, but Agentic AI can help align them.

A great example is chronic disease management. Today, many health plans and providers can identify patients with diabetes, heart failure, COPD, or kidney disease who are at high risk for complications. The challenge isn't identifying them—it's engaging them at the right time with the right intervention. An Agentic AI system can continuously monitor clinical, claims, pharmacy, and behavioral signals, identify rising-risk patients, recommend personalized interventions, coordinate outreach, and adapt engagement strategies based on patient response. Preventing even a small percentage of avoidable hospitalizations can generate significant savings while improving quality of life for patients.

Another powerful use case is closing care gaps. Healthcare organizations spend enormous resources trying to improve preventive screenings, medication adherence, and quality measures. Agentic AI can identify members with open care gaps, determine the most effective intervention for each individual, coordinate outreach across multiple channels, and continuously monitor completion rates. This improves patient outcomes while helping providers and health plans perform better on quality programs such as HEDIS, Medicare Star Ratings, and value-based contracts.

A third area is reducing avoidable emergency room visits and hospital admissions. Many patients show warning signs weeks or months before an acute event occurs. Agentic AI can detect these signals early, prioritize patients based on risk and likelihood of intervention success, and recommend actions that prevent deterioration. The result is fewer emergency visits, fewer inpatient admissions, and better patient outcomes.

Behavioral health is another area where Agentic AI can have a major impact. Mental health conditions often go undetected until a crisis occurs. By combining claims, medication, utilization, and social determinant signals, AI agents can identify members who may benefit from earlier intervention and connect them to appropriate resources before their condition escalates.

For self-funded employers, Agentic AI can help optimize high-cost populations by identifying members likely to experience catastrophic events, predicting specialty drug utilization, recommending care management interventions, and evaluating which point solutions are most likely to deliver measurable ROI. This allows employers to lower healthcare spend while improving employee health outcomes.

Perhaps the most exciting use case is personalized healthcare navigation. Healthcare is incredibly complex, and patients are often left to navigate it on their own, often relying on fragmented information from multiple sources. Agentic AI can serve as an intelligent health guide, continuously analyzing an individual's clinical history, benefits, care gaps, and preferences while also incorporating the latest evidence from peer-reviewed medical literature, clinical guidelines, and best practices. It can help patients understand their treatment options, identify high-quality providers, manage chronic conditions, stay current with preventive care, and make more informed healthcare decisions. By delivering personalized, evidence-based guidance at the right moment, Agentic AI has the potential to improve outcomes, enhance the patient experience, and lower overall healthcare costs.

Ultimately, the biggest opportunity isn't automating individual tasks. It's creating an intelligent healthcare orchestration layer that continuously identifies risks, recommends interventions, coordinates actions, and learns from outcomes. That's where we can begin to address the trillion-dollar waste problem in healthcare while simultaneously delivering more personalized, higher-quality care.

 

4. What challenges must healthcare organizations overcome to successfully deploy AI agents at scale, including trust, governance, and regulatory concerns?

 

While the potential of Agentic AI is enormous, healthcare organizations must overcome several important challenges before AI agents can be deployed at scale.

The first challenge is trust. Healthcare decisions directly impact people's lives, so clinicians, care managers, and patients need confidence in the recommendations being made. AI cannot operate as a black box. Organizations need transparency into why a recommendation was generated, what data was used, what evidence supports it, and the level of confidence behind the recommendation. Building explainable and auditable AI systems will be critical for adoption.

The second challenge is governance. Unlike traditional software, AI agents can make dynamic decisions based on changing information. Organizations need clear governance frameworks that define what actions agents can take autonomously, what actions require human approval, how performance is monitored, and how unintended consequences are identified and addressed. Human oversight must remain an integral part of the process, particularly for high-impact clinical decisions.

Data quality and interoperability are also major hurdles. AI agents are only as effective as the data they receive. Healthcare data is often fragmented across claims systems, EHRs, pharmacy platforms, laboratory systems, and other sources. Organizations must establish strong data foundations and interoperability strategies to ensure agents have a complete and accurate understanding of the patient and their healthcare journey.

Privacy, security, and regulatory compliance are equally important. Healthcare organizations must ensure that AI systems operate within HIPAA requirements, protect sensitive patient information, maintain robust cybersecurity controls, and provide appropriate safeguards around data access and use. As AI becomes more autonomous, organizations will need even stronger controls to ensure patient privacy and security are never compromised.

Another important challenge is validating clinical accuracy and safety. Healthcare organizations cannot simply deploy an AI model and assume it will perform as expected. Agentic AI systems must be continuously monitored, tested, and evaluated against clinical outcomes, quality measures, and operational metrics. Just as new medications undergo rigorous evaluation, AI systems need ongoing validation to ensure they remain accurate, reliable, and safe.

From a regulatory perspective, we're still in the early stages. Regulatory frameworks are evolving rapidly as policymakers seek to balance innovation with patient safety. Organizations should design their AI strategies with the expectation that requirements around transparency, accountability, documentation, and oversight will continue to mature over time.

That said, I believe the biggest challenge is not technology—it's organizational readiness. Successful adoption requires healthcare organizations to rethink workflows, redefine how humans and AI collaborate, and build a culture that embraces AI as a tool for augmenting human expertise rather than replacing it. The organizations that succeed will be those that combine strong governance, responsible AI practices, and human oversight with a clear focus on improving patient outcomes.

Ultimately, trust will determine the pace of adoption. If clinicians, patients, regulators, and healthcare leaders trust that AI agents are transparent, safe, evidence-based, and acting in the best interest of patients, Agentic AI has the potential to transform healthcare at a scale we've never seen before.

 

5. What are the key takeaways attendees can expect from your session on "Agentic AI in Healthcare: Eliminating $1T in Waste. Lowering Costs. Improving Patient Outcomes"?

 

Ans: Attendees can expect three key takeaways from my session.

First, they'll gain a practical understanding of what Agentic AI actually is and why it represents a significant shift beyond traditional AI, automation, chatbots, and copilots. There is a lot of hype in the market today, and I want attendees to leave with a clear framework for understanding how AI agents can collaborate, reason, and take coordinated action to achieve healthcare outcomes.

Second, we'll explore how Agentic AI can address some of healthcare's most pressing challenges. Despite spending more than any other nation on healthcare, the U.S. continues to struggle with rising costs, fragmented care, clinician burnout, preventable hospitalizations, and poor consumer experiences. I'll share real-world use cases showing how AI agents can help close care gaps, improve chronic disease management, reduce avoidable utilization, support behavioral health, optimize healthcare navigation, and improve quality outcomes while lowering costs.

Third, attendees will learn how to move from experimentation to implementation. Many organizations are currently piloting AI solutions, but few have a roadmap for deploying Agentic AI at scale. We'll discuss the foundational capabilities required for success, including data integration, governance, trust, security, clinical validation, human oversight, and responsible AI practices.

Most importantly, I want attendees to leave with a broader perspective on the role of Agentic AI in healthcare. Today, much of the focus is on using AI for clinical documentation, prior authorization, customer service, and reducing administrative burden. These are valuable use cases that can improve efficiency and reduce costs, but they only scratch the surface of what's possible.

The much larger opportunity lies in using Agentic AI to transform healthcare decision-making. By continuously analyzing data from multiple sources. AI agents can generate personalized, actionable recommendations for patients, providers, care managers, and health plans. Instead of simply automating existing workflows, Agentic AI can help identify risks earlier, recommend the next best action, close care gaps, guide consumers through complex healthcare decisions, and ultimately improve outcomes while lowering costs. That's where I believe the true transformational potential of Agentic AI exists. If we get this right, we can meaningfully reduce the more than $1 trillion in waste that exists within the healthcare system while delivering better outcomes and experiences for everyone involved.

 

 

Bio:

 

AI executive with expertise in enterprise AI, Generative AI, and scalable AI platforms. Former leader at NVIDIA and Microsoft, they have driven AI initiatives generating $1B+ in growth and launched 30+ AI-powered products across healthcare and enterprise technology.

 

Abstract

The U.S. healthcare system loses nearly $1 trillion annually to inefficiencies—fragmented data, delayed interventions, administrative burden, and a lack of actionable intelligence. Despite the explosion of data from claims, electronic health records, and social determinants of health (SDoH), care teams still struggle to answer two critical questions: Who should we prioritize? and What intervention will actually work?

This is where Agentic AI represents a fundamental shift.

Unlike traditional black-box AI, which often produces predictions without transparency or actionability, Agentic AI systems don’t just predict—they reason, decide, and act. They continuously monitor patient data streams, identify rising risks before they become costly events, and autonomously recommend—or even trigger—evidence-based interventions.

In this talk, we explore how Agentic AI can:

• Surface invisible rising risks (before hospitalizations occur) 

• Close care gaps in real time using HEDIS, STAR, and RAF-driven insights 

• Integrate fragmented data across clinical, claims, and SDoH domains 

• Reduce administrative burden by automating decision workflows 

• Deliver personalized, evidence-based care pathways at scale 

• Drive measurable cost savings and deliver 10x ROI through proactive, targeted interventions 

We will also examine real-world applications—from risk stratification and proactive outreach to cost-saving recommendations and care coordination—highlighting how Agentic AI transforms healthcare from a reactive system into a proactive, intelligent ecosystem.

Ultimately, the future of healthcare is not just about more data—it’s about intelligent agents that turn data into outcomes. By addressing inefficiencies at their root, Agentic AI has the potential to eliminate waste, improve equity, and fundamentally reshape how care is delivered.

 

📅 Join us on June 25 for the AI in Healthcare Conference and hear directly from industry experts on the latest advancements in Agentic AI, Generative AI, Clinical AI, Healthcare Data Platforms, Digital Health, and more.

 

🎟️ Register today and reserve your spot! https://www.eventbrite.com/e/global-ai-in-healthcare-virtual-conference-tickets-1988295787223