Relatient

A scheduling platform whose core is a rules engine, with AI layered on as channels. The company describes Dash as 'AI-powered, rules-driven automation': an engine encoding each provider's preferences, locations and visit types decides what can be booked, and it runs online self scheduling, centralized call center scheduling, messaging, digital registration and financial clearance without any model.

Dash Voice AI is a conversational agent onto that engine that books, cancels, reschedules and confirms appointments, routes calls by intent and routes requests such as prescription refills and medical records to clinical teams; AI-enabled chat does the same in text. The voice and chat agents depend on models; the scheduling decisions depend on the rules.

The oversight structure is described and one part is missing. The voice agent books, cancels, reschedules and confirms appointments without staff involvement and writes changes straight to the record or practice management system, but only within the same rules engine the human schedulers use, and the health system can limit what the agent may verify or change by appointment type, provider or location, preventing changes to high-complexity or time-sensitive appointments.

Calls that need a person are handed to staff with the caller's details carried over, and the company reports 70 percent of calls resolved without human intervention across 1 million inbound calls. What is missing is the condition that routes a call to a person, and what the agent does when a caller describing a routine request mentions a worrying symptom; the agent now takes clinical requests such as refills and medical records and routes them to care teams.

The workflow is described clearly and the models are not. Public material explains the rules engine shared by human and automated scheduling, the configurable safeguards on what the voice agent may change, real-time write-back, the option to finish booking through a secure text link, intent-based call routing, and routing of clinical requests to care-team tools.

It publishes an automation rate (70 percent of calls resolved without human intervention) and a patient satisfaction score (4.85). No speech or language model is named, no accuracy figure is given for recognizing what a caller wants or for extracting a clinical request, no language coverage is stated, and no versioning or update practice is published.

The architecture is described and no provider is named. A voice agent and chat sit over a rules engine, write to record and practice management systems, send secure text links, and route clinical requests into care-team workflow tools; the voice channel therefore passes caller audio through speech recognition and language understanding before anything reaches the scheduling rules. No speech, language model, telephony, messaging or hosting provider is named, and no subprocessor list is published. Since July 2026 clinical, medication and billing requests travel the same path.

Vendor-reported percentages and named customers, with no method. The voice AI page reports 1 million inbound calls answered, 70 percent resolved without human intervention, a 14 percent increase in new patient appointments and a 4.85 patient satisfaction rating; the home page reports 97 percent fewer cancellations with self-scheduled appointments, a 10 percent increase in physician utilization and 60 percent of calls deflected. None names the customer, period or baseline behind it.

Named customers speak in quotes: Tennessee Orthopaedic Alliance describes a 30 percent pickup in appointment volume within three months, a support services director describes the AI handling 56 percent of calls, and UMP and Complete Health are named. The company reports more than 50,000 providers and over 150 million appointments a year and won Best in KLAS in 2024; scale and ranking are recorded but are not evidence of benefit.

Security assurances are published; the questions AI raises are not answered. The security page describes HIPAA aligned safeguards, a SOC 2 Type 2 examination and HITRUST i1 certification across call center, online and voice agent scheduling.

The company's HIPAA Authorization page asks patients to authorize Relatient to use and disclose their protected health information for purposes including enabling and customizing their use of its services, beyond what it may do as a business associate of their providers. Nothing published states whether calls are recorded, how long audio and transcripts are kept, whether calls or patient data train the voice or chat models, or de-identification. Since July 2026 the voice agent captures clinical, medication and billing requests, which is a different kind of information from an appointment slot.

Relatient states that it is a business associate and publishes a document addressed to patients. The company's HIPAA Authorization page, linked from the site footer, explains that companies like Relatient that assist covered entities are business associates, that Relatient collects, uses and exchanges health information with patients' providers as part of its services, and asks patients to authorize its use and disclosure of their protected health information for listed purposes; information not covered by HIPAA falls under its privacy policy, which covers the website.

The security page describes HIPAA aligned safeguards alongside HITRUST i1 and SOC 2 Type 2. The business associate agreement itself is not published, so its terms and whether it binds subcontractors cannot be read before contracting.

Two recognized attestations are named as the company's own, with their levels, and no artifact to open. The security page, linked from the site footer, states a SOC 2 Type 2 examination and HITRUST i1 certification for the scheduling platform, covering call center, online and voice agent scheduling, and a 2026 press release announces HITRUST recertification. The HITRUST i1 is the one-year implemented level, below the risk-based r2. No trust center, report request route, audit period, penetration testing statement or subprocessor list is published.

No device claim is made and the core work is administrative: booking, rescheduling, canceling and confirming appointments, routing calls by intent, registration and financial clearance. Since July 2026 the voice agent also captures patients' clinical, medication and billing requests, such as prescription refills and medical records, and creates a task for the right care team in its existing workflow tools.

That is message handling rather than a clinical judgment, and nothing published says the agent answers, ranks or triages those requests. The exposure is the call nobody planned for: a patient describing a worrying symptom while booking, and nothing describes what the agent does then. Automated outbound calls and texts carry telephone consumer protection consent obligations.

Nothing is published on how the voice and chat agents are evaluated, monitored or tested: no accuracy figure for understanding callers, no error or misroute rate, no results by group and no governance framework. Customers get analytics to track what works and fine-tune call flows, which is operational reporting rather than evaluation of the models. The exposure is access.

The voice agent answers the general scheduling line for practices and health systems serving more than 50,000 providers, and speech recognition performs unevenly across accent, dialect, age and speech difference; a caller the agent repeatedly fails to understand does not get booked or is sent to the queue the automation exists to shorten, and leaves no record of why. Language coverage is not stated.

Mechanisms exist to trace and contain an error, and nothing stands behind the output. Because the voice agent books through the same rules engine as human schedulers, a disputed booking traces to a rule someone configured rather than to model behavior; health systems can restrict what the agent may change; transferred calls carry the caller's details; and the care team gets call history and context after the call.

The 70 percent of calls resolved without human intervention is an automation rate, not an error rate, and no misbooking, misroute or extraction accuracy figure is published. No warranty, remedy or service credit attached to a wrong booking or a lost clinical request is described.

Named two-way integrations with major record systems, one of them verifiable in the record vendor's own marketplace. Relatient reads live schedules and writes appointment changes to the record or practice management system in real time, and routes clinical requests into care teams' existing workflow tools. The site carries dedicated integration pages for athenahealth, ModMed, Oracle Health, Greenway Health, eClinicalWorks, NextGen, Epic and Veradigm, plus scheduling APIs.

In April 2026 athenahealth named Relatient a Preferred Solution Partner, available through the athenahealth Marketplace, with athenahealth citing a decade of results for mutual customers, which is production evidence from the record vendor itself. Deployments across mixed record estates are described. The other named integrations are verifiable only on Relatient's own pages.

Nothing published says where the platform runs or where patient data rests. The security page describes certifications and safeguards, the voice AI page describes real-time write-back, and the website privacy policy covers site visitors; none names a hosting provider, region, residency commitment, tenancy model or subprocessor. A hosted service is implied, and the voice channel adds call audio, transcripts and clinical request content to what is stored.

No price is published, and the shape of the offer is discoverable. The Dash platform is sold through a demo and RFP route as modules (intelligent scheduling, voice AI agent, AI-enabled chat, personalized communications, digital registration, financial clearance and scheduling APIs) to hospitals and health systems, specialty groups and practices, and federally qualified health centers. It is also available through the athenahealth Marketplace.

Relatient automates patient access administration: online self scheduling, centralized call center scheduling and a voice agent that books, reschedules, cancels and confirms appointments, call routing by intent, digital registration, financial clearance, and routing of patients' refill, records and billing requests to the right team.

Coverage is named by buyer: hospitals and health systems, specialty groups and practices, federally qualified health centers, and since 2026 behavioral health, with orthopedics, dermatology and reproductive health among the specialties represented on its customer advisory board. Specialty handling comes through each provider's scheduling rules rather than specialty-specific models.

Validation sits behind part of it, with Tennessee Orthopaedic Alliance, UMP and Complete Health named and a customer base of more than 50,000 providers. Coverage is United States only and stops at patient access.