AI Knows What You Should Eat

What’s Inside

* How Viome analyzes RNA to personalize what people should eat

* Why to avoid even “healthy” foods like broccoli or turmeric

* What founders can learn from Viome’s use of AI as an interface

Guru Banavar has worked at the frontier of AI for years.

He helped lead IBM Watson before pivoting into healthcare. Now as co-founder and CTO of Viome, he’s applying that AI background to a new challenge: personal health.

Banavar and I talked about how Viome uses RNA data to power a personalized approach to nutrition, disease prevention and health optimization. In his case, the analysis flagged a biological imbalance triggered by too much turmeric in his diet.

AI Meets Biology: RNA, Not DNA

Viome’s platform centers on metatranscriptomics, a process that captures RNA expression from saliva, stool and blood samples. That data — now over a million samples strong — runs through Viome’s proprietary AI engine to detect biological patterns and deliver actionable recommendations.

“Even the healthiest foods,” Banavar said, “can actually be not good for you.”

That line captures the shift in thinking. Viome doesn’t treat AI as a chatbot. It treats it as an analytical layer built to surface biological insights that conventional tools or generalized advice often miss.

Future of Healthcare Is at Home

What stood out was how accessible Banavar sees this becoming. “We believe the future of healthcare is really at home,” he said. “You take the tests, we do the analysis, and you get recommendations that fit your biology — not a general trend or diet.”

In his case, the platform flagged an issue with bile acid pathways tied to turmeric. He also noted that some users show inflammation responses to broccoli, depending on their sulfide pathways, a level of dietary personalization that’s hard to achieve without molecular data.

AI as Infrastructure, Not Interface

Banavar also offered a reframe of how Viome uses large language models.

“You can think of the LLM as the UI,” he said. “The real IP is in the molecular analysis, the recommendation engine, and the bioinformatics we’ve built.”

For founders and operators, that takeaway is especially sharp: the insight lives below the interface. The real value comes from the infrastructure.

Viome offers a clear example of how AI, when rooted in biology and delivered at scale, can offer something far more useful than general advice: actionable personal agency.

Don’t Miss Out on More Content

This episode originally aired on The AI Cognitive Shift, a show I co-host with AiNews.com. You can watch the original video version on YouTube or subscribe to The Venture Variety Show here on Substack, as well as my own YouTube channel.

And if this conversation sparked ideas or reframed something for you, feel free to share it.

Get full access to The Venture Lens at theventurelens.substack.com/subscribe