Compare

An AI receptionist answers your phone. We run your front door.

Several of these companies are well funded and genuinely good at the phone. Naming them and still winning the row is the point.

Capability comparison across DocAssistants, Prosper AI, Hello Patient, and Hyro
CapabilityDocAssistantsProsper AIHello PatientHyro
Owns the websiteYes — we build and run itNo — voice and workflow, not the public siteNo — phone and schedulingNo — conversational layer on existing properties
Writes to long-tail EMRsPractice Fusion and eClinicalWorks live; certified FHIR R4 as a pathBroad EHR coverage, including large-system vendorsScheduling integrations; coverage varies by practiceEnterprise integrations; strongest in health systems
Human release before publishYes — digest-bound gate. Nothing patient-facing leaves without a named personCall review and exception routing when confidence is low — not a publish gateStaff can take over a live call; not a pre-publish releaseHandoff to live agents; not a content-digest gate
Full audit logHash-chained record of every tool call and state changeCall reports, transcripts, and QA flagsCall logs and scheduling historyConversation analytics and QA

A distinction that gets blurred

Escalation is not a release gate.

Escalating a call when a model’s confidence is low is exception routing. It is not a human approval gate. A gate means nothing goes out at all until a person releases it, bound to the exact content approved. If the bytes change after that release, the publish is refused.

Judge it on your own line.