Autonomous AI that adjusts medications, orders labs, and documents in the EHR — without a scheduled visit — just received FDA clearance. The billing and coding infrastructure for what it creates doesn't fully exist yet.
Oversubscribed seed round — backed by Mayo Clinic, Eli Lilly, the American Diabetes Association, Cathay Innovation, Pear VC, Polaris Partners, and Section 32.
UpDoc announced today that it has received FDA clearance for the first Software as a Medical Device (SaMD) built on patient-facing large language models. The Palo Alto company simultaneously disclosed initial deployments at Cleveland Clinic, Allegheny Health Network, and UCSF Health.
The platform integrates directly with a health system's EHR and operates on behalf of a licensed clinician between patient appointments. The company's published example: a Type 2 diabetes patient whose blood glucose is drifting. UpDoc identifies the trend, initiates an insulin titration within physician-approved parameters, triggers a follow-up lab order, and documents the entire intervention in the EHR — without requiring the patient to schedule an appointment.
The key distinction from administrative AI is that the platform is performing clinical actions. Not scheduling, not eligibility checks, not prior auth submission. It is executing care. The clinician sets parameters; the AI executes within them.
Autonomous between-visit care coordination is not a hypothetical billing scenario. The codes exist. The question is whether health systems have the infrastructure to capture them at the volume this platform would generate.
Relevant code categories for autonomous between-visit interventions:
The challenge is documentation and attribution. Traditional CCM billing requires time-stamped documentation of the clinical interaction, a documented care plan, and provider attestation. When the clinical interaction is an AI agent executing an insulin titration at 2 AM, the billing team needs to know it happened, who owns the documentation, and whether the EHR entry constitutes a billable event under the relevant payer's policy.
UpDoc's platform triggers lab orders autonomously. Lab orders — even within physician-approved parameters — are subject to prior authorization by many commercial payers, particularly for high-volume chronic disease management panels. The automation that eliminates the scheduling friction doesn't eliminate the PA requirement on the downstream lab order.
Health systems deploying UpDoc at scale need to understand their payer mix's PA requirements for the lab panels their UpDoc protocols will trigger, and ensure the platform either integrates with their PA workflow or routes flagged orders for human review before transmission.
Does a CCM billing note generated by an AI agent satisfy the attestation requirements of CMS and commercial payers? This is currently unresolved. Expect initial denials followed by policy clarification over the next 12–18 months as volume scales.
The $18M seed included strategic investors whose involvement functions as market signals as much as capital. Mayo Clinic's participation implies clinical validation and a potential deployment pathway. Eli Lilly's involvement — as the manufacturer of the leading GLP-1 and insulin products — signals direct commercial interest in AI-assisted titration and adherence management at population scale. The American Diabetes Association as an investor is directionally the same signal.
UpDoc's FDA clearance confirms regulators will approve patient-facing agentic AI for clinical use. The revenue opportunity from autonomous between-visit care is real — CCM, RPM, PCM codes exist and reimburse. But documentation standards, payer policies, and audit exposure haven't caught up to what the platform can generate at scale.
Source: UpDoc press release, June 25, 2026
Payer policy changes, denial trends, vendor intel — delivered every morning. Free.
Reaching RCM decision-makers? Advertise with RevCycleAI →