Steer Health

Everything Steer sells is agent delivered and there is no non AI business underneath: it is not an EHR, a payments platform or a services firm with AI added. Note for scoping rather than for the grade: documentation is one module among many in a platform whose commercial center of gravity is patient acquisition and revenue growth, so a buyer evaluating Tara as a standalone scribe is buying into a much wider system.

Wide autonomous reach with no published limits. Agents answer inbound calls, schedule, manage referrals, handle insurance verification, create care plans and write actions back into the EHR unattended, described by the vendor as completed work rather than call deflection.

For the documentation module specifically, no review gate, confidence threshold, acceptance rate or escalation path is published, and care plan creation in particular is a clinical artifact rather than an administrative one. The 98 percent accuracy figure is offered in place of an oversight description, which is not the same thing.

A 98 percent note accuracy figure is published with no methodology, reference standard, denominator or date, and the site offers no model card, named models or evaluation protocol. Nor does anything published say whose model it is, whether the agents run on proprietary models or a general purpose foundation model with a healthcare layer.

No party in the chain is named, but the architecture is described. The vendor states that every agent runs on SteerBrain, which it describes as a purpose built reasoning system with clinical rules, payer logic and operational rails integrated at the model level, trained on more than 1.8 billion healthcare interactions, and explicitly not a wrapper on a public large language model or a white label product.

That is a claim about the model layer rather than a disclosure of it. It does not say whether a foundation model sits underneath the reasoning system, who hosts it, or which parties handle audio and call recordings. No subprocessor list is published.

The content moving through that unnamed chain is broad. Agents handle inbound and outbound calls, eligibility and payer work, triage, clinical documentation and outreach to lapsed patients, and the homepage voice demo states that calls are recorded to improve the agent. Much of what the platform captures is patient speech outside any clinical encounter, and the vendor markets one shared patient thread across every agent, which makes retention architectural rather than incidental.

Named customer evidence is substantial, and what it measures is commercial. Steer reports deployment across 472 or more live locations and 100 or more health systems, names Prime Healthcare with more than $24 million in incremental net revenue across 38 facilities, and cites specific results at named organizations including a 72 percent lift in online new patient conversions and a 47 percent lift in surgical conversion.

Every outcome measured is commercial rather than clinical or documentation quality: demand created, inquiries booked, slots filled, patients returned, revenue captured. None is independently verified, and none shows whether the notes are any good.

The vendor publishes nothing on audio or transcript retention, deidentification or training use. That gap is wide here because of how much the platform touches.

Seven agents span inbound telephony, insurance verification, prior authorization, triage, ambient documentation, discharge follow up and lapsed patient recall, across a claimed four hundred and seventy two locations. Each produces a different class of content and several produce recorded patient speech outside any clinical encounter. A triage call, a discharge check in and a recall campaign contact are all patient interactions held by the vendor, and none of them is a visit.

Two categories deserve their own answer rather than being folded into a general retention question.

Recorded patient speech from the voice agents is the larger volume by some distance, given ten thousand daily check ins are claimed, and it involves patients who did not attend a visit and may never become one. And the recall agent operates on lapsed patients, which means the platform retains and acts on data about people who have stopped engaging with the practice. What is held about a patient who never returns, and for how long, is not stated.

The vendor's stated memory design makes this sharper rather than softer. It markets that the platform remembers every patient so its agents can act with continuity, which is a real product benefit and also means retention is architectural rather than incidental.

The vendor states that a business associate agreement is standard on every deployment, alongside health privacy compliance and SOC 2 Type II certification. Standard on every deployment is the formulation a buyer wants: the agreement travels with the product rather than with the negotiation, which settles whether smaller customers or pilots receive one.

The published legal surface is thin by comparison. The terms of use a buyer can open are general platform terms rather than healthcare terms. They reserve the vendor's right to use aggregated and de identified data for analytics and improvements, and say nothing about business associate obligations. The agreement itself is not published, so its scope cannot be read.

Scope is the real question because of the platform's breadth. Its agents answer inbound calls, check eligibility, work prior authorization and denials, triage and schedule, draft clinical notes and run outreach to lapsed patients. Eligibility and authorization work sends patient information onward to payers, and outreach reaches patients directly. Nothing published states whether one agreement covers every agent or names the subprocessors at each step.

SOC 2 Type II certification is stated first party, on the vendor's about page and on its comparison page, where it is presented with HIPAA compliance and a business associate agreement as standard on every deployment. Type II tests whether controls operated effectively across a period, which is the stronger report type.

No trust portal, report date, audit period, auditor or scope statement is published. An older security page that also claimed HITRUST compliance now redirects to the homepage, so that claim is no longer current on the vendor's site, and nothing there says whether any HITRUST certification exists.

Scope is the open question. This is a multi agent platform reaching telephony, eligibility and payer workflows, triage, documentation and patient outreach, with a shared patient memory underneath. Which of those systems the report covers cannot be inferred from a certification named on a marketing page.

No clearance is claimed, and ambient documentation needs none. The positioning question is triage, not documentation.

The vendor presents one agent as a triage nurse and scheduler, markets routing orthopedic and spine demand into routine, urgent and surgical, and its terms list clinical triage automation among its services. Routing a patient by acuity in a call where no clinician is present is the kind of output device regulation is concerned with, because the patient acts on it immediately rather than a professional reviewing its basis first. No regulatory position is stated for that agent.

The vendor does draw the clinical line in writing. Its published limits state that clinicians sign every note and that complex clinical judgment stays with the clinician. That keeps documentation clearly outside device territory. It does not say where triage sits relative to that line.

The incentive structure comes from the pricing model rather than the marketing. Steer offers performance aligned pricing, stating that clients can pay based on outcomes delivered rather than a flat subscription, and one of the outcomes it markets for its documentation module is a 15 percent revenue lift from surfaced missed billing and coding opportunities.

Where that pricing applies, the vendor's own compensation rises as the practice's coding intensity rises, which is a sharper alignment question than a revenue lift advertised alongside a fixed fee. Non contingent subscription pricing does not reward coding more; outcome pricing tied to revenue can. There is no claim of removing the human from coding: the vendor states that complex surgical sub specialty coding needs a coder in the loop, so the concern is alignment, not autonomy. No fairness statement, subgroup analysis or accent disclosure is published.

The vendor publishes a section of limitations. It states that clinicians sign every note, that complex clinical judgment stays with the clinician, that prior authorization is an assisted capability rather than full automation, and that complex surgical sub specialty coding needs a coder in the loop. Review before commit on notes and a person in the loop on complex coding are real mechanisms for challenging an output.

Nothing stands behind the output, and the published terms point the other way. The terms of use cap the vendor's total liability at the fees paid in the prior twelve months, exclude consequential damages, and require the customer to indemnify the vendor. No warranty, remediation commitment or error rate for any agent is published.

The limits do not address the agents that act without a clinician present. A triage agent routes patients, and a voice agent books appointments and verifies coverage in live calls; no error rate, escalation rule or correction route is published for either. The vendor says its impact report discloses methodology for every metric, and that report may hold the error rates.

Strong integration claims, asserted rather than verified. Steer describes writing directly into Epic, Oracle Health and athenahealth plus a claimed 50 or more other systems, with bidirectional synchronization named specifically for Epic and Cerner, pre built integrations, and go live in three to four weeks from contract including EHR integration, agent configuration and staff training. Writing completed actions back rather than deflecting calls is the meaningful distinction. No integration architecture, certification program membership or named production reference for the write back specifically is published.

The vendor's own material names no hosting region, residency option or subprocessor list.

The integration position is stronger. The vendor describes a certified bidirectional interface across the record systems a customer already runs, naming the two largest and more than thirty others, with a dedicated interoperability agent writing every agent action back into the record in real time. Bidirectional write back through a permissioned interface is the right architecture.

The write back scope is also unusually broad. Every agent action being written into the record means scheduling, insurance verification, prior authorization activity, triage outcomes, documentation and outreach contacts all land in the chart from an automated source. It is not stated what identity those writes carry in the audit log, whether they are distinguishable from staff actions, or what an institution can restrict per agent.

One reported detail would matter if confirmed. A third party account covering several similar platforms describes a zero day retention agreement with a named model provider. If this vendor holds one it would answer the subprocessor question directly, since zero retention at the model provider is the specific commitment that makes an external processor acceptable. The attribution is unclear, and the vendor's own material does not confirm it.

The mechanism is disclosed clearly and the number is not. Steer states that pricing is performance based, with clients paying on outcomes delivered rather than a flat software subscription, customized to organization size and services deployed, and quoted through sales. Disclosing the model openly says more about the commercial relationship than a rate card would. But the vendor gives no figure, floor, or definition of which outcomes trigger payment, and outcome definitions are precisely where a performance based contract is won or lost.

Broad across care settings, and reaching into cash pay medicine. Coverage spans hospitals and health systems, medical groups, surgical and medical specialties with operating room block time and referral workflows, and emergency departments where it targets left without being seen rates and door to provider times.

It also runs a dedicated functional medicine vertical built around membership and program based business models, recurring revenue, entitlement tracking and advanced laboratory panel workflows, a cash pay segment. No specialty count is published, and language support is English only.