For more than two decades, Ratnakar Lavu’s career in technology focused on retailers including Macy’s and Bloomingdale’s. As his career matured, he moved into the C-suite to serve as chief technology officer at department store Kohl’s and global chief digital information office at sportswear giant
Nike.
But in early 2024, Lavu took a big leap into an entirely new industry. He became CDIO at health benefits provider
Elevance Health, which ranks
No. 20 on the Fortune 500 with $197.6 billion in operating revenue in 2025, a 13% increase from the prior year. Lavu says that although the industries are vastly different, both can benefit from investments in technology intended to make it easier to serve customers.
“My past experience has been with large consumer brands; we’re always focused on leveraging technology to transform brands and create delightful, personalized customer experiences,” says Lavu. “At Elevance Health, I’m trying to do the same thing.”
To be sure, leveraging technology to bring “delight” to the complex and expensive healthcare system is a far loftier challenge than selling sneakers or designer purses. American satisfaction with U.S. healthcare costs is at the
lowest level Gallup has recorded since 2001, with only 16% now saying they are satisfied, down from the all-time high of 30% in the first year of the COVID-19 pandemic. Nearly one in four say the U.S. healthcare system is “in a state of crisis,” while 47% believe it has “major problems,” according to a poll published in December.
Lavu says he believes that AI can help simplify the system and improve health outcomes, and he focuses his investments on three areas, all of which have some exposure to generative AI. For the company’s 45.2 million medical members, there’s a ChatGPT-like experience built into Elevance Health’s Sydney mobile app, where users can ask questions in natural language, such as, “I need knee surgery. Am I covered and if so, how much will my co-pay be?” Sydney can answer those questions. The chat feature can also recommend providers who are regionally close to the member and are also in network, which is more cost effective than an out-of-network health practitioner.
The second pillar centers on Elevance Health’s health plans, which include its Anthem Blue Cross and Blue Shield, Wellpoint, and Carelon brands. Lavu has deployed AI for prior authorization, a process that requires a doctor to obtain approval for a health plan before providing treatment or writing a prescription. AI is automating some of those requests, though humans are still in the mix for some evaluations.
Elevance Health has also built an internally developed AI chatbot called Spark, which leverages multiple large language models to handle document analysis and other common corporate tasks. Elevance Health says more than 60,000 associates use Spark and in the first half of 2025, 10 million messages were sent through the chatbot.
The company is also in the process of deploying an enterprise ChatGPT license across the company and is working with the AI startup to give tens of thousands of associates an opportunity to earn certifications in AI fluency, which includes responsible AI, prompt engineering, and more advanced AI-enabled work.
“We’re really dedicating our time and effort to make sure all our employees are upskilled,” says Lavu.
Lavu is still in the process of making meaningful headway on agentic AI, though he says this form of the technology is especially promising within healthcare, as it can handle more complex tasks across benefits and claims. He has been exploring agentic AI use cases across all three of Elevance Health’s priority pillars and expects to scale them more aggressively in 2026.
What Lavu is now focused on is setting up the proper guardrails to clearly define what tasks these agents can perform and, just as importantly, what they shouldn’t do. He wants to very precisely train them on only the right data needed to perform their tasks. And employees will always remain in the mix.
“A human can actually oversee multiple agents,” says Lavu. “But there has to be some oversight, because we want to ensure that the agent does not deviate from its task that it’s been assigned.”
John KellSend thoughts or suggestions to CIO Intelligence here.
