Care delivery · Clinical AI
Akido
A Los Angeles company that owns clinics and builds the AI that runs the visit inside them. An assistant sits with the patient, a language model works out the diagnosis, a physician signs it. The evidence that this matches a doctor-led visit has not been published.
The visit that does not look like one
A patient checks in at an Akido cardiology office in the Inland Empire. A medical assistant in his mid twenties, DeAndre Siringoringo, takes her to an exam room, takes vitals, reviews her medications and asks what brought her in. He is warm and unhurried. Intake runs about thirty minutes.
What she does not see is the software listening. It is called ScopeAI. While Siringoringo talks, it feeds him the next question, assembles the history, ranks possible diagnoses and drafts a treatment plan, logging what it ruled out and why.
When intake ends, the system pings a physician, someone like the Pomona cardiologist Haritha Alla. She reads the note, the leading diagnosis, the alternatives and the plan, then walks in and confirms or changes it. Her corrections feed the training loop.
Siringoringo told MIT Technology Review he tells patients an AI system will be listening to gather information for their doctor. He does not tell them it also proposes the diagnosis.
The arithmetic Akido argues from
Every Akido document rests on one pair of numbers: the United States needs more than 3 billion doctor visits a year and roughly 825 million are available. The AAMC projects a shortfall of up to 86,000 physicians by 2036. Nobody trains their way out of that in a decade.
So Akido inverted the visit. Instead of a doctor gathering information and then reasoning, an assistant gathers it under machine direction and the doctor reasons over a finished draft. Chief executive Prashant Samant says this lets a doctor see four to five times as many patients.
The first people it reaches are on Medicaid, and not only for reasons of mission. Medicaid lets a doctor approve ScopeAI plans asynchronously. Many commercial insurers still require the doctor to speak with the patient first, so the model works best where patients have the fewest alternatives. Samant says that gap is a feature of how insurance works rather than a choice.
From a data lab to a medical group
Akido started in 2015 inside the USC Keck School of Medicine’s Digital Health Lab, founded by Samant, Jared Goodner and Sanjit Mahanti, and spun out through Y Combinator. For five years it was a data company, stitching clinical and social service records into what it describes as more than 10 million longitudinal case studies.
Selling software into other clinics capped the impact, so in 2022 Akido bought its way into care delivery through Chaparral Medical Group in Pomona, then in December 2023 took a controlling stake in Rhode Island Primary Care Physicians Corp, a network of 150 plus practitioners. Terms were undisclosed. Goodner calls the result a “living-learning laboratory where technology and operations evolve together.” It is also why Akido can redesign the visit without anyone’s permission.
The story changes shape
In July 2026 the company dropped “Labs” from its name, launched a patient-facing brand called Akido Medical, and said ScopeAI now reaches 100 clinics, virtual care and field medicine in California, New York City and Rhode Island.
The language moved with it. In September 2025 the chief technology officer described the goal as pulling the doctor out of the visit. Ten months later the company was selling “a diagnostic head start” and “a better, more human experience with a real doctor spending more time with each patient,” care “delivered by humans, not chatbots.”
In between, in January 2026, Akido’s own release described a company that “builds AI-powered doctors and delivers clinician-supervised care through its own multi-specialty provider network.” That went out the same month California’s AB 489 took effect, barring AI systems from implying a user is getting care from a licensed professional unless licensed oversight is present.
What is proven, and what is still claimed
| Evidence | What the record shows | Source type |
|---|---|---|
| FDA status | No 510(k) and no PMA record as of September 22, 2026. Akido’s position, per Forbes: because clinicians make all final decisions, the system is clinical decision support and exempt from device regulation. | Public record |
| What regulators say | Neither the FDA nor the Medical Board of California would tell MIT Technology Review whether ScopeAI was on solid legal footing. Harvard Law’s Glenn Cohen said a system acting as a doctor in a box would likely need FDA approval and could run afoul of licensure law. | Independent |
| Accuracy bar | Before a new specialty goes live, Akido requires the correct diagnosis to appear in ScopeAI’s top three at least 92 percent of the time on historical data. Self-set, retrospective, unpublished. | Company-stated |
| TAMSS benchmark | July 2026: tested against 203,000 encounters under a Total Addressable Medical Situation Space framework Akido built itself. No methodology, results or audit released. | Company-stated |
| Prospective comparison | None. MIT Technology Review reported Akido had not run studies comparing ScopeAI appointments with in-person or telehealth ones. No study naming Akido or ScopeAI appears on ClinicalTrials.gov. | Not found |
| Peer-reviewed work | Two PubMed papers carry an Akido Labs affiliation, both predating ScopeAI: a COVID-19 self-assessment tool (2020) and an emergency data capture tool (2022). Neither concerns diagnosis. | Public record |
| Federal grants | None found at NIH or NSF. Street medicine runs on Medi-Cal reimbursement. | Not found |
| Street medicine results | The numbers move between tellings. Fierce Healthcare, January 2026: two-thirds of patients seen on day one, 70 percent six-month retention, ED visits down 40 percent in some areas. Akido’s own 2026 white paper: 53 percent seen on day one, 63 percent six-month retention, ED use down 55 percent against a national rate. | Company figures differ |
| Series B size | Announced as $60 million on May 15, 2025. The Form D filed the next day reports $84,254,689 sold to 94 investors, first sale March 11, 2025. | Differs from filing |
Read plainly: Akido has built something real, and its scale makes the missing evidence matter more, not less. It has a defensible regulatory theory, a working business and satisfied patients. What it lacks is one published comparison showing a ScopeAI visit produces an outcome as good as a doctor-led visit. Its accuracy bar is self-set, its benchmark self-built, its clinical numbers drifting between releases. That is not evidence of harm. It is the absence of the evidence that would settle the question.
What to watch
- A prospective study comparing ScopeAI visits with physician-led visits. Nothing is registered, and it is the result that would answer the automation bias question.
- Whether TAMSS is published or audited, or stays a marketing term.
- Enforcement of California’s AB 489 and AB 3030, and whether patients are told that the system listening also proposes their diagnosis.
- A new Form D. Nothing has been filed under Akido’s CIK since May 2025, despite growth from 240 providers to over 1,000.
- Whether commercial insurers drop the requirement that a doctor speak to the patient, which is what confines this model to Medicaid.
In their words
“Our focus is really on what we can do to pull the doctor out of the visit.”
Jared Goodner, CTO, MIT Technology Review, 2025 · Independent
“Scope spans 26 different specialties and performs at an attending level of investigation and diagnosis in those areas.”
Prashant Samant, CEO, The Heart of Healthcare podcast, 2026 · Interview
“It allows me to be in 10 places at once.”
Steven Hochman, Akido street medicine physician, MIT Technology Review, 2025 · Independent
“With Scope, I can see double the patients in the same period of time. This means my patients have much more access to me and I can add new patients, which is unheard of in the current moment with physician shortages.”
Haritha Alla, Akido cardiologist, Forbes, 2025 · Independent
“I worry that it might predispose you to sort of nodding along in a way that you might not if you were actually in the room watching this happen.”
Emma Pierson, UC Berkeley, on automation bias, MIT Technology Review, 2025 · Independent
“We designed ScopeAI specifically to reduce bias by proactively countering blind spots that can influence medical decisions, which historically lean heavily on physician intuition and personal experience.”
Akido spokesperson, MIT Technology Review, 2025 · Company-stated
“Laws must evolve to reflect the role of AI in care delivery. Physicians shouldn’t be left as the sole fall guys for machine-driven errors.”
Dana Corriel, MD, SoMeDocs, 2025 · Independent
Related companies
Sources
- Public recordForm D, Akido Labs, Inc., CIK 0002063139
- Public recordFDA 510(k), PMA and ClinicalTrials.gov: no Akido record
- Public recordCOVID-19 Web-Based Patient Self-assessment Tool, PMID 32542495
- Public recordElectronic Data Capture Tool for Public Health Emergencies, PMID 35679114
- Public recordGenAI Notification Requirements, AB 3030
- IndependentThis medical startup uses LLMs to run appointments and make diagnoses
- IndependentAkido Labs And The Coming Shift To AI-Enabled Medicine
- IndependentAkido Labs launches AI-powered street medicine program in Bay Area
- IndependentA Growing Akido Calls Itself First AI-Native Health System
- IndependentAkido Makes Rhode Island Play
- IndependentWho’s Liable When AI Gets It Wrong?
- InterviewWhat It Takes To Scale Care With AI
- IndependentAI in Healthcare Faces New Guardrails Under AB 489
- CompanyAkido Raises $60 Million Series B
- CompanyAkido Announces the First AI-Native Health System in the US
- CompanyAkido Labs Announces Strategic Leadership Appointments
- CompanyStreet Smart white paper and Scope AI page
Profile researched and written by Healthcare Discovery. Last updated September 29, 2026.
