UpDoc’s prescription-only SaMD has FDA clearance for type 2 diabetes patients. [Image courtesy of UpDoc]
The news caught my eye, as I’m always trying to learn more about innovative, first-of-their-kind medical devices, especially when there are lessons other device developers could learn that would help them succeed. And artificial intelligence is an area where I’m looking for medtech use cases and real results, because there’s so much hype around AI in medtech and beyond, particularly around large language models (LLMs)
UpDoc describes itself as “an enterprise-ready clinical AI company developing physician-grade agentic AI.” It’s got former FDA Commissioner Dr. Robert Califf as an advisor and says it has raised $18 million in seed funding from investors including the American Diabetes Association, Eli Lilly and Co. and Mayo Clinic.
The 2026 MDO Medtech Startups Special Report: Technical tips and actionable advice for device startups
“At the core of the platform is UpDoc Clinical Intelligence, a proprietary orchestration and safety architecture that unifies clinical guidelines, longitudinal patient context, physician governance and EHR data to coordinate and execute real-world clinical care safely — at scale,” the company said in a news release. “The platform automates interventions that traditionally required a physician’s direct, hands-on involvement with tasks such as adjusting medications, facilitating lab orders, coordinating with care teams and more.”
UpDoc’s says its type 2 diabetes SaMD is “designed to support doctors, not replace them.” [Image courtesy of UpDoc]
I hadn’t heard of UpDoc before, and I was curious why a first-of-its-kind medical device using LLMs would go through the 510(k) regulatory pathway instead of the de novo route for novel devices. I reached out to the company for their regulatory documents and a representative confirmed this 510(k) clearance from December 2025 was for the SaMD.
Related: Philips asks physicians and patients how they’re using AI
What UpDoc’s FDA 510(k) clearance filings say
The regulatory filing lists the indication for prescription-only use as follows:
“UpDoc is a software as a medical device (SaMD) intended to provide medication management for patients aged 18 years and older who have been diagnosed with type 2 diabetes. UpDoc provides patients with insulin treatment plan instructions based on a healthcare provider (HCP)-specified treatment plan.”
Deep dive: Diabetes device developers turn their sights to type 2
The indication for use includes two user-interactive software components: a user interface for patients and another for healthcare providers.
The patient user interface is the UpDoc mobile application, which is “intended for use by patients with type 2 diabetes as an aid in optimizing insulin management. Patients use the mobile application to log blood glucose, meal, symptom, and medication adherence data, and receive treatment plan instructions. Data may be entered manually or reported via voice or text-based interactions. The application may also receive blood glucose data via a Bluetooth-enabled glucometer or continuous glucose monitor.”
The healthcare provider user interface is the UpDoc web portal, which is “intended for use by trained healthcare providers to configure and manage the patient-specific insulin treatment plan. This includes insulin dosing instructions (type, starting and maximum doses, adjustment algorithm, and blood glucose targets) and safety protocols to address non-emergency hypoglycemia, hyperglycemia, and related symptoms. Insulin instructions are computed in UpDoc’s cloud-based application based on the HCP-defined treatment parameters.”
A third software component is UpDoc’s cloud-based application, which includes the UpDoc Agent conversation service and a clinical service. Insulin instructions are computed in that cloud-based app based on treatment parameters defined by the healthcare provider.
MDO Contribution: FDA rules, regulations and resources for artificial intelligence in medical devices
UpDoc’s 510(k) predicate device is the d-Nav Insulin Management Program, which provided insulin dosing recommendations. The software was developed by Hygieia and FDA-cleared in 2019 before Hygieia filed for bankruptcy in 2024.
UpDoc says its use of voice and chat interfaces for data entry is a key difference from that previous software, along with “additional safety features related to symptom reporting and system lock behavior.”
“UpDoc does not interpret or diagnose symptoms,” the company says in the FDA filing. “It only allows patients to log the presence or absence of symptoms, and considers this data when providing patients with treatment plan instructions. The presence of symptoms that are specified in a pre-determined treatment protocol may result in a change in insulin dose or a system lock with instructions to the patient to seek medical attention. The presence of symptoms that are not specified in a pre-determined treatment protocol will result in a system lock and instructions to the patient to seek medical attention. These features enhance patient safety by preventing insulin dose increases in the presence of newly reported symptoms, thereby reducing the risk of inappropriate dosing.”
The SaMD uses what UpDoc describes as “conversational data collection” to let a patient “enter data through voice or chat interfaces in addition to manual input. Within the UpDoc Patient App, patients may speak with the UpDoc Agent via voice or chat to input their blood glucose values, adherence data, meal data, and symptoms, and receive new treatment plan instructions.”
Beyond offering insulin dose instructions, UpDoc can also tell patients they should “eat something sweet, contact their provider, or seek immediate medical attention in situations in which the patient reports low or high blood glucose values or specific symptoms. All instructions are based on the predefined treatment plan configured by the HCP.”
UpDoc co-founder and CEO Dr. Sharif Vakili [Photo courtesy of UpDoc]
“We fundamentally believe that clinical AI should be held to the highest standard,” Vakili said in the release announcing the product. “This FDA clearance marks a historic milestone for the field and the beginning of a new era in clinical care, where clinicians will be able to deploy clinical AI agents to help complete tasks autonomously on their behalf, much like engineers deploy coding agents today. This breakthrough also has the potential to make primary care viable again by directly addressing some of the healthcare system’s most pressing challenges: patient access, rising system costs and clinician burnout.”
I’ve asked UpDoc for more details on their use of AI and LLMs specifically, their product development journey and their regulatory approach, and I’ll share more information if and when it’s available.
Related: Medical robotics and AI advice from a J&J DePuy Synthes leader
