Using AI to reinvent care delivery at Novant Health

Novant Health appointed Vijay Sankararaman as its first chief AI officer last year to reshape care delivery using artificial intelligence. The health system has deployed a virtual care AI agent for clinically accurate medical answers, predictive AI for proactive patient outreach, and generative AI to draft insurance claim appeals, aiming to improve patient experience and operational efficiency.

In healthcare, the pressure to improve outcomes while reducing costs has never been greater. Yet many organizations are still applying AI to existing systems rather than using it to fundamentally change them. At Novant Health, that shift is being driven in part by establishing the chief AI officer role last year, which Vijay Sankararaman has held since then. He uses AI to reshape how care is accessed, delivered, and supported across the enterprise, all while being a catalyst to rethink how the organization operates end to end. How is AI helping Novant Health achieve its vison? Our cause is to provide a remarkable experience for our patients, to make it easier, more valuable, and more meaningful for them to access the healthcare we provide 24/7 across primary care, specialty, ambulatory, emergency, and pharmacy. But we live in a changing world with economic headwinds for patients and their healthcare teams, so transformation for us is about reimagining the patient experience, ease of use for clinicians, and value-add from our operations. That’s our AI lens. The mantra is simple. AI isn’t put on top of a stack of technology and processes. It’s a vehicle that allows us to challenge the status quo and ask if there’s a better way, how would we construct a hospital if we started today, and what’s my ideal working environment as a nurse. We take the persona of the team member and patient, and reimagine their work with the power of predictive, generative, and agentic AI. What are some examples of this strategy in action on the patient side? We all love to ask our phones if nagging knee pain is chronic or acute, but the information we receive isn’t always valid. Novant Health created a virtual care AI agent that answers medical questions safely and clinically accurate, using your health records within secure firewalls. The agent gives the patient better information, with correct patient data, and the privacy choice not to put their personal health information into ChatGPT. We’ve also launched a proactive outreach program for high-risk patients who don’t know they’re at risk. We use predictive AI to identify them based on their health factors, and conversational AI to reach out and encourage them to schedule diagnostic scans. Early detection, after all, saves lives. What are you doing on the operational side with AI? Healthcare is complex with a variety of reasons for payers to deny claims, and for providers to challenge. But in the middle are patients who are stuck. We use gen AI to draft claim appeals, which allows us to represent the patient’s interest, give clarity to the claims challenge, and raise patient satisfaction. We’re also using AI to improve patient advocacy in a very meaningful way. Whether it’s cancer outreach or patient advocacy, healthcare requires significant human capital, not just at Novant, but in any 24/7 provider environment. AI doesn’t sleep. This CAIO role didn’t exist a year ago. How do you define it? The mandate of the chief AI officer is to enact and orchestrate the transformation that leverages operational excellence, innovation, and technology while keeping the patient at the center. My team and I are designing the new choreography that runs across the entire organization. Our CIO and his organization are essential partners, leading the core infrastructure across data, cyber, and ERP/EHR, while my organization complements that work by designing the system of the future. We have product leaders, AI engineers, and software and platform engineers who build homegrown AI solutions alongside the hyper-scalers and domain-specific language model partners, all of whom are positioned to leverage AI. What is the advantage of a CAIO as a peer to IT leadership? The peer seat allows AI leaders authentically to interrogate how the organization is functioning. What are our real success metrics? What are we doing to achieve them? How can automation and innovation get us to that North Star faster? Every business leader is in pursuit of a goal, and every pursuit has friction. Traditionally, when a business leader brings that friction to IT, they present it as a problem, a solution, and a directive to use this technology to fix the problem. With AI, our business leaders start with a problem and an interest in consuming data differently. They want help thinking through the problem, they aren’t ordering a solution. That shift is significant. The people-process-technology framework is now influenced by AI, which provides a level of context not available before. AI is a context creator, rather than a content creator. We haven’t had that before. What advice would you give to healthcare CEOs right now? Three things keep healthcare CEOs awake at night. The complexity and cost of healthcare are rising, prompting people to avoid going at all. The nurse population — arguably the most important and most populous in healthcare — is dwindling. And the cost of staying in business is a challenge, between Medicare, Medicaid, and everything else. As a CEO, I’d start with literacy. The CEO and leadership team must understand what AI can do. Break those opportunities into two buckets of excellence and enrichment. Excellence is how you engage with patients and run your operations. Enrichment is augmentation in that AI is a companion to the clinician, noting very early-stage detection. This is happening at Novant Health now. Then start to innovate, in both physical and digital spaces. We’re actively leveraging precision robotic surgery and launching an AI solution for diabetic retinopathy, a common high-risk condition for people who don’t get regular eye exams, or something they don’t know they have. A primary care physician, augmented by AI, can now detect it. That’s where physical meets digital. Whether they decide to hire a chief AI officer or not isn’t the point. The most important question is are you thinking about AI with a viable lens. Do you have a purposed organization and leader fully empowered to challenge the status quo. If you don’t, that’s where to start. When should companies hire a CAIO as a peer to the CIO? There’s no one set formula. In some companies, the CAIO sits above the CIO and at others, below. At Novant, our organizations are healthy peers. What clicks with the DNA of your company is the right thing to do. My CIO and his organization are true partners to our transformation team. We’re completely aligned, and we advance the mission together. If you’re not in a position where a CIO and CAIO can advance your position, then don’t hire one. It’s what makes sense right now for us, and like so much with AI, it’ll change over time.