Artificial intelligence startup Keebler Health raised $16 million in series A funding as it continues to build AI-powered infrastructure for value-based care.
Flare Capital Partners led the series A round, with participation from Sands Capital, Tau Ventures and existing investors—including Freestyle Capital, Underdog Labs and MBX Capital. Additional investors in the round include Everywhere Ventures, New Stack Ventures, Tweener Fund, Aviano Ventures and Hustle Fund, according to a press release.
The startup has raised $23 million to date, according to Isaac Park, its CEO and co-founder.
Risk adjustment remains constrained by fragmented medical data, with most meaningful patient information residing outside of coded fields, according to executives.
Keebler Health uses an LLM-native risk-adjustment platform that is built to process unstructured clinical documentation. The tool creates accurate and complete Hierarchical Conditional Category (HCC) coding that provides clinicians with actionable insights at the point of care, executives said.
“As we were exploring that technology front, we realized that risk adjustment would actually be a really great place to be able to start affecting this kind of technology change that we were leveraging,” Park told Fierce Healthcare in an exclusive interview about the series A funding round.
The core challenge in risk adjustment is the gap between what is documented and what is captured in coded fields, Keebler Health executives said. Approximately 80% of healthcare information is unstructured, residing in narratives, imaging reports and discharge summaries rather than in coded fields. And even structured records are often incomplete: A study published in the Journal of the American Medical Informatics Association found that only 59.4% of chronic conditions are consistently captured across electronic health record sources. The result is systematic gaps in risk capture and reimbursement accuracy.
“What stood out to us about Keebler is how clearly the team is executing against a long-standing limitation in healthcare data,” said Ian Chiang, partner at Flare Capital Partners, in a statement. “They’ve built a platform that aligns with how clinical information is actually documented and are already demonstrating the ability to turn that into meaningful, measurable results. We believe this is a critical capability for organizations operating in value-based care, and we’re excited to support the team as they scale.”
Keebler Health plans to use the funding to expand its team, continue commercial growth and support infrastructure for value-based care organizations nationwide. It will also support the company’s expansion into adjacent use cases, including compliance and audit workflows, such as AI-enabled RADV audit readiness.
“Finding at scale, that real, unstructured clinical story on these patients has been that core value driver for us in risk adjustment for these value-based care organizations,” Park said.
The company raised $1.8 million in a pre-seed funding round led by New Stack Ventures and $6 million in a seed round led by Freestyle Capital, Park told Fierce Healthcare.
Launched in 2023, the Durham, N.C.-based company was founded by Park, Andrew Stickney, Kevin Hill, Ph.D and founding Chief Medical Officer Terrell Bacchus, M.D. Park said the group “adjacently” ran into issues that got them thinking about clinical narratives.
“We landed on this real opportunity [of] finding that unstructured clinical narrative and processing it in such a way that we can use software at scale to pull out evidence that we had found in those clinical areas,” Park said.
Park said one of the challenges around risk adjustment has been the lack of an outside tool aside from “traditional natural language processing tooling,” which has since changed with the advent of LLM technology.
“Really, for the first time in history, you're able to use software at scale to read parts and understand unstructured clinical narrative,” Park said. “We have built a tool that allows us to go back longitudinally as much unstructured clinical data as we can get our hands on and really understand what’s happening to the patient narrative based off of the notes, not necessarily just the claims.”