Keebler Health

High-Level Overview

Keebler Health is an AI-native risk adjustment platform designed for healthcare providers and payers managing Medicare and Medicaid patients in value-based care (VBC) models. It uses transformer AI models and large language models (LLMs) to analyze full patient charts—including structured data, unstructured notes, handwriting, scans, faxes, and PDFs—identifying chronic conditions, HCC codes, and risk adjustment opportunities with source-linked evidence for audit-proof accuracy.[1][3][6] The platform serves ACOs, primary care groups, home health/hospice providers, and health plans, solving the problem of fragmented data and missed diagnoses that lead to inaccurate RAF scores, revenue loss, and suboptimal patient care by providing clinician insights, recommended next steps, and scalable automation that cuts costs by up to 70% compared to traditional methods.[1][2][3][4] Founded in 2023 and based in Durham, North Carolina, Keebler has raised $7.8 million in seed funding, including a $6 million round in 2024 from investors like Freestyle Capital and MBX Capital, fueling team growth and product acceleration amid rising VBC demand.[1][2]

Origin Story

Keebler Health was founded in 2023 by CEO Isaac Park, COO Andrew Stickney, and CTO Kevin Hill in Durham, North Carolina.[1][5] The idea emerged from Park and Stickney's prior co-investment in a medical company, where they observed that 5-7% of procedure costs went to inefficient billing and coding—separate from providers and insurers—highlighting "symptoms" of poor handoffs between clinical documentation and coding teams.[5] This inspired a "Grammarly for medical coding" co-pilot using proprietary AI to detect inconsistencies, omissions, and hidden complexities in notes, reducing denial rates and boosting compliance without fraud.[5] Early validation came from seeing the AI surface overlooked risks that improved patient care, confirming their vision for an LLM-native platform empowering entire organizations to handle risk adjustment across all patients, not just high-risk samples.[1][2]

Core Differentiators

Keebler stands out in risk adjustment through AI-driven innovation tailored for VBC:

• LLM-Native and Transformer AI: Trained on all clinical data types (handwriting, scans, longitudinal notes), enabling full population analysis without sampling, unlike traditional code-scanning or expert-only methods.[1][3][6]
• Comprehensive Data Ingestion: Functions as a "HIE of HIEs," unifying EHRs, faxes, PDFs, claims, and consults into gap-free, audit-ready patient records in days, not months, with FHIR and HIE integrations handling "dirty data" without normalization.[3][4][6]
• Explainability and Trust: Every HCC code or RAF opportunity includes confidence scores, source evidence, and counterfactuals, fostering clinician adoption and reducing burden with embedded workflow insights.[1][3][6]
• Cost and Scalability: Automates to cut costs by 70%, scales nationally, supports v28 HCC codesets, and provides clinical decision support like predictive signals and pre-visit planning.[1][2][3]
• Workflow Fit: Boosts CDI, coding, and audit teams with actionable, non-noisy signals, even for non-EHR-owning groups.[3][4][6]

Role in the Broader Tech Landscape

Keebler rides the VBC wave, where providers bear financial risk for Medicare/Medicaid outcomes, amplified by CMS rule changes demanding broader risk capture beyond high-risk sampling.[2] Timing aligns with LLM advancements enabling organization-wide analysis of unstructured data—previously limited to experts—amid market forces like rising ACO adoption, audit pressures, and the need for accurate RAF scores to protect shared savings.[1][2][4] It influences the ecosystem by democratizing risk adjustment, reducing reliance on offshore vendors/spreadsheets, and improving care via clinician tools, positioning it against players like Lumeris and Lightbeam while accelerating VBC scalability for health systems.[1][2]

Quick Take & Future Outlook

Keebler is primed to expand its platform with recent funding, targeting team growth, roadmap acceleration, and penetration into ACOs/health plans as VBC mandates intensify.[2] Trends like AI clinical reasoning, FHIR ubiquity, and whole-patient data hubs will propel it, potentially evolving into a full VBC intelligence layer with predictive care beyond risk adjustment. Its clinician-trusted, scalable edge could redefine coding copilots, leveling up providers in a risk-bearing world much like its AI boosts overlooked insights today.[1][2][6]