Remote Monitoring & Chronic Care
U

UpDoc

UpDoc runs clinical AI agents that manage chronic conditions between visits under a physician's prescription. A clinician prescribes an individualized treatment plan tied to an agent, through the record system or the provider portal, setting the insulin type, starting and maximum doses, the adjustment algorithm, glucose targets and safety protocols for nonemergency low and high blood sugar. The patient connects existing monitoring devices, logs readings and symptoms, and talks with the agent by call or message. A conversation service built on a large language model handles that exchange, while a separate clinical service applies the clinician's rules with deterministic logic and returns the treatment action for the conversation service to relay. Clinical notes are recorded back into the record, and clinicians follow measurements, adherence and treatment progress in the portal, where every intervention traces to the prescribed plan.

The FDA cleared UpDoc V1.0 under K253281 in December 2025 for prescription insulin management in adults with type 2 diabetes, covering the patient app, the provider portal and the cloud service that calculates instructions. The company launched the cleared product in June 2026 with Cleveland Clinic, Allegheny Health Network and UCSF Health among its first deployments, and the wider platform also supports hypertension. In September 2026 ARPA-H picked UpDoc to lead its ADVOCATE program, an award of up to $9.2 million to build heart failure agents with support from Microsoft, OpenAI and NVIDIA.

Health systems and plans contract with UpDoc, and patients get access through a participating provider. Cofounder and chief executive Sharif Vakili is a primary care physician at Stanford Medicine, where his 2023 trial of a voice intervention for basal insulin preceded the product. UpDoc raised an $18 million seed round from investors including the American Diabetes Association, Eli Lilly and Mayo Clinic.

AI Health Index verifiedNovember 19, 2026
Compare UpDoc with other vendors
Founded
—
Headquarters
—
Website
updoc.ai/
Categories
remote-monitoring, medication-safety-and-prescribing
Indexed Products
UpDoc Clinical AI, UpDoc Insulin Management, UpDoc Provider Portal, UpDoc Patient App
Buyer Segments
Health systems, Primary care and diabetes programs, Health plans, Clinicians managing chronic conditions
Assessment

Capability Axes

An AI Health Index grade measures what a buyer can verify from public sources on the date shown. It is not a rating of how good the product is. A vendor can build an excellent system and grade low on an axis because it publishes nothing an outsider can check. How grades read

AI Capability
BB on AI CentralityThe model is the engine of a core module. The platform carries other value, but this capability does not exist without it.
Regulatory Filing

The patient agent runs the conversations that collect readings, symptoms and medication adherence, then relays instructions and follow up prompts, which is the recurring contact between visits the platform exists to deliver. The provider portal organizes treatment plans and lets clinicians follow changes over time.

For the FDA cleared insulin function, the clinical service computes instructions from the clinician's plan with deterministic logic, while the conversation model handles the patient side. The ARPA-H program pushes the same split toward heart failure, with agents adjusting medications and escalating to clinicians. Without the agents the service has no way to reach patients between visits.

BB on Autonomy and Oversight ModelThe oversight structure is described and one part is missing, commonly the threshold at which the system stops or what happens after it is wrong.
Regulatory Filing

A clinician prescribes the service and sets an individualized treatment plan and safety protocols, and the software then conveys instructions between visits under that plan by call or message. The clinician monitors progress and changes treatment settings in the provider portal, patients are told to contact their clinician with concerns, and the cleared use covers nonemergency management only. Every intervention traces to the prescribed plan, and health systems can review that record.

Once enrolled, the patient receives each instruction without a clinician signing off on it, and the platform's line that complications are caught before they turn into hospitalizations comes with no stated escalation threshold.

BB on Model and Technology TransparencyThe approach or the suppliers are named without the version and update discipline behind them. Naming a supplier is the entry to this band both here and on Model Supply Chain Disclosure, which ask different questions of the same disclosure: who receives the data, and what produces the output.
Regulatory Filing

UpDoc Clinical Intelligence coordinates the patient app, the clinician portal and two cloud services. The conversation layer integrates a large language model for the patient interaction, and the clinical service uses rules the provider defines to decide how to act under an insulin plan, a design that keeps dosing deterministic. The FDA summary identifies the device version and the types of change the authorized change plan covers.

The commercial conversation model, the model version in each deployment and its training sources are not identified. Clinicians define the treatment plan without configuring the conversation model, so what they control is the clinical rule set rather than the language layer.

CC on Model Supply Chain DisclosureThe architecture is described and no model provider is named. Naming a hosting provider alone does not lift a record out of this band. Record the host in the note, because it matters for residency and breach scope, and grade on the model layer, which is the question this axis is named for.
Vendor Published

The marketed insulin function separates the conversation model from the clinical service, which uses deterministic treatment logic, and both the cloud service and the patient app sit inside the cleared device architecture. The ARPA-H ADVOCATE program names Microsoft, OpenAI and NVIDIA as supporting partners for the heart failure agents it will build over 39 months.

That syndicate describes future development and does not identify which third party model, if any, processes patient conversations in deployed accounts today. The architecture shows where an outside dependency could sit, but no production model supplier, pinned model version or customer choice of supplier is named.

BB on Clinical and Operational EvidenceNamed deployments with dated outcome figures and enough method to test them, or published research short of independent validation.
Peer Reviewed Publication

In a 2023 randomized study of 32 adults across four primary care clinics, published in JAMA Network Open, participants using a predecessor voice intervention reached an optimal basal insulin dose in a median of 15 days versus more than 56 days with usual care, with adherence of 82.9% against 50.2% over eight weeks. Those results are the source of the platform's headline figures, a threefold improvement and a 65% rise in adherence.

The founders ran that study at Stanford before the current language model integration, and the FDA review of the cleared product describes nonclinical and human factors validation. Outcomes from the 2026 health system rollout have not been reported.

CC on AI Safety and PHI StewardshipGeneral assurances of privacy and security that do not answer the questions artificial intelligence raises: what is retained, what reaches a model, and what happens to it there.
Vendor Published

A provider sets safeguards for the insulin plan, including maximum doses and protocols for nonemergency low and high blood sugar, and the clinical service bases each instruction on those settings. Separating dose calculation from conversation keeps the language model away from the treatment decision itself.

The patient app processes conditions, medications, vital signs and conversations with the care assistant, and the privacy policy allows disclosure to supporting service providers and gives patients a way to make privacy requests. The patient terms set no retention schedule for conversation records and no separate commitment on whether those conversations are used to train models.

Regulatory and Compliance
CC on HIPAA and BAA PostureA status is stated but not supported: compliance is claimed, or business associate status stated, without the underlying document, including where the only privacy notice published covers the website rather than the service that handles patients. The mirror case grades here as well: a substantive privacy document that reaches the service, with no statement of business associate status anywhere and no scope position taken. One of the two elements the band above asks for is present, which is a grade below it and not a grade at the floor.
Vendor Published

The trust center lists HIPAA compliance, and health systems enroll patients in a service whose app and cloud components handle treatment details, measurements, symptoms and adherence. The privacy policy applies to the app and uses information to deliver the service, coordinate care and support software operations, with disclosure to authorized service providers.

The online terms set no account tier that carries a business associate agreement, and they do not say which model suppliers or service partners an institution's agreement would cover. With patient conversations running through a language model, the agreement's reach to that supplier is the open question for a health system, and it sits in the institutional contract.

BB on Security Certifications and Trust CenterA recognized certification is named in the vendor own material without the artifact, or with a scope or renewal question the buyer has to raise. A recognized certification here is an audit of how the vendor operates security, such as SOC 2, HITRUST or ISO/IEC 27001. A certification has a scope and a clock, and both are part of this grade. A certification the vendor attaches to its data center, hosting facility or cloud provider, rather than presenting as its own, is the host's and does not count here.
Vendor Published

The trust center lists SOC 2 Type II and HIPAA and gives a contact for security questions. The mobile app and provider portal work with a cloud service that stores or processes information collected during treatment, including patient conversations with the agent. The FDA review covered cybersecurity submissions for the insulin device, which apply to the cleared version.

The attestation is named, but the report itself, its audit period, the cloud components in scope and controls specific to the patient conversation service are not open to review. The platform pages link only privacy and terms, so the trust center carries the whole security posture.

AA on FDA and Regulatory StatusThe regulatory position is unambiguous and verifiable: a clearance or authorization identifiable in the public databases, with the version and indication it actually covers. A product outside device regulation reaches this band when the vendor publishes why, naming the exclusion it relies on, and markets nothing past it.
Regulatory Filing

The FDA cleared UpDoc V1.0, version 01-00-US, under K253281 on December 23, 2025, against a drug dose calculator predicate. Its prescription indication covers insulin management in adults with type 2 diabetes following a healthcare provider's treatment plan, and the clearance describes a patient app, a provider portal and a cloud application with separate conversation and clinical services.

The company launched the cleared functionality in June 2026 and calls it the first FDA cleared agentic clinical AI platform with patient facing language models. Hypertension and heart failure programs sit outside this 510(k) indication, and the ARPA-H plan routes heart failure through investigational device exemption trials.

BB on AI Governance and Bias DisclosureA governance framework with named process behind it, such as certification to an artificial intelligence management standard, or material written for a customer own review committee to evaluate the product with. A framework covering the vendor whole portfolio counts when its scope reaches this product; what it cannot supply is the product specific evaluation the band above asks for. Published results within the one group where the performance of this product is most at risk also sit here: evaluation on the question this axis asks, short of the comparison across groups the band above requires. A medical device clearance or certification, and the quality system behind it, does not reach this band on its own; it is graded under regulatory status. A regulator reviewed plan governing how the model may change does reach it.
Regulatory Filing

The FDA authorized a predetermined change control plan for specified updates to the insulin device, covering certain changes to default values, clinical definitions, supported insulin references, interface behavior and input methods. The plan keeps the clinical service's deterministic logic and sets a process for assessing and validating each change. Providers control each patient's treatment settings, and a Clinical AI Charter states the company's principles for the agents.

The authorization does not reach unrestricted changes to the conversation model, and no breakdown of clinical outcomes by patient subgroup is given for the marketed version, a gap for a population as varied as adults on insulin.

CC on AI Liability and RecourseMechanisms exist that let someone challenge an output, such as audit trails, source traceability or review before commit, with nothing standing behind the output and no route for the harmed party.
Vendor Published

The prescribing team keeps authority to set the plan and to respond when a patient raises a clinical concern. Patient instructions send health concerns to the treating clinician, and a provider can review interventions and update the plan, with each intervention traceable to the prescribed treatment plan. The privacy policy gives a separate contact for access and correction requests about personal information, which covers the data record rather than a clinical event.

Clinical incident handling, turnaround times for an investigation and remedies when a software instruction is disputed are not described, so escalation duties fall to each healthcare organization's own care program.

Integration and Deployment
BB on EHR and Interoperability DepthNamed systems with read access or one directional writing, or standards support with named deployments behind it.
Vendor Published

Physicians prescribe individualized treatment plans tied to an agent from their record system, and clinical notes are recorded back into the record automatically, which sets out a defined two way workflow. Connected monitoring devices add data alongside what patients report. The FDA filing describes the provider portal as the clinician interface used to configure and monitor the cleared device.

Allegheny Health Network called record integration a future step at the June 2026 launch, and the ARPA-H program plans record integrated deployments at health system scale. No record system is named, and none of the initial deployments shows both directions running in production yet.

CC on Deployment Model and Data ResidencyA single hosted option with location implied rather than committed.
Vendor Published

The provider portal runs in a browser and patients use a mobile app that talks to the cloud service, which holds both the conversation layer and the clinical calculation for the cleared insulin function. The prescription device is offered in the United States through participating institutions, and the US indication sets the market where that version may be prescribed.

The platform is built for enterprise health system environments and their compliance requirements, and it runs around the clock between visits. No cloud regions are named for storage or model processing, there is no residency commitment a customer can select, and there is no option to run the service inside a health system's own environment.

Commercial
DD on Commercial TransparencyNothing a buyer can establish before a sales conversation. A published pricing claim contradicted by evidence also grades here.
Vendor Published

Health systems and payers arrange an institutional deployment through sales, and at a participating organization the patient's clinician prescribes access and manages the treatment plan. The mobile app downloads at no charge, but the care service behind it needs enrollment through a provider. The pitch to payers promises durable outcomes at a fraction of the cost of traditional care models, without a figure behind it.

No rate per patient or per site is listed, and neither is a minimum commitment or an implementation fee for connecting the record system, so what the software and its role in the care program cost comes only from a contract.

BB on Setting and Specialty CoverageCoverage is named with validation behind part of it.
Vendor Published

The first deployments named at the June 2026 launch were Cleveland Clinic, Allegheny Health Network and UCSF Health, with health plans as the other buyer. The patient uses the service at home while the provider follows the treatment course in the portal. The FDA indication covers adults with type 2 diabetes who need prescription insulin management, and the platform dashboard also runs hypertension programs.

The ARPA-H award extends the same model to heart failure care, drawing on connected scales and blood pressure cuffs. The basal insulin trial ran at four primary care clinics in one academic center, so the evidence comes from a specific outpatient setting, with limited coverage of the other disease programs the platform now offers.

Alternatives

Alternatives to UpDoc

Other products that do the same kind of work as UpDoc, and what each one does.

AI enabled chronic care management for older adults, run in partnership with health systems: AI agents monitor daily vitals from connected devices in patients' homes and handle medication titration, lifestyle coaching, alert triage, care coordination and symptom assessment, working with…

AliveCor makes ECG devices and the AI that interprets them.

Delfina combines maternal health software with remote monitoring and support for pregnant and postpartum patients. Its app collects symptoms, mood, weight, blood pressure, and glucose information, while a clinician dashboard brings those measurements into the care team's workflow.

Verily Health is a precision health company that became independent of Alphabet in April 2026, when it was renamed; Alphabet remains a significant minority investor after a $300 million round led by Series X Capital.

Nanowear weaves the sensors into clothing.

Vivalink sells the sensing and data layer that other organizations build monitoring products on.

Commercial

Pricing

Vendor-published figures are labeled as such. Figures labeled “Estimated” are derived from third-party sources and have not been confirmed by the vendor.

Entry Price Pricing Basis BAA Tier Implementation Source
Contact for pricing
Institutional agreement through sales No business associate agreement tier is stated. HIPAA compliance is listed in the trust center No implementation or record integration fee is listed. Vendor Published

UpDoc sells through health systems and payers, and patients get prescription access through participating providers. The app download is free. Subscription rates and record integration charges are not listed.