Prosper AI

Conversational voice agents are the entire product. There is no non AI version of an agent that navigates a payer IVR, holds, and speaks with a representative.

The oversight structure is described with one part missing. Prosper's FAQ publishes the escalation conditions: when the agent is unsure it transfers to staff, offers a callback or takes a voicemail; language suggesting an emergency stops the automation, plays a safety script and directs the caller to emergency care or on call staff; a distressed caller is handed to a person; complex or clinical requests are handed off with context.

For prior authorization it gathers details and initiates next steps with the practice's team. Every call gets automated QA. What is missing is what happens after an agent obtains or records wrong benefits information. The FAQ line that patients 'never get stuck in loops' describes routing and is recorded as a claim.

The supplier is named without version discipline. Prosper names OpenAI as its underlying model provider and describes the quality architecture: every call scored by automated QA for accuracy and compliance, with recordings and transcripts passed back to the customer. No model version, accuracy figure for benefits capture or change practice is published.

The model provider is named with its terms; the rest of the chain is not. Prosper names OpenAI and states a zero day retention agreement with it, so PHI sent to the model is not stored or used for training there. It lists the other kinds of party that touch PHI (speech to text, text to speech, the telephony carrier) without naming them, and publishes no subprocessor list.

Anonymized outcomes and scale, with no method. Prosper reports a Northeast OB/GYN group with about half of scheduling calls automated and a Northeast GI group of more than 100 providers with more than half of front desk scheduling and waitlist volume handled by AI within weeks, and says re engagement campaigns can lift appointments 20 percent or more. The measures that matter for payer calls, benefits accuracy and prior authorization completion, are not published with a method.

Categorical commitments with part of the schedule. Prosper's HIPAA guide states a zero day retention agreement with OpenAI, its model provider, so protected health information sent to the model is not retained there or used for training, and daily backups kept under a 30 day retention policy. What is not stated is how long call recordings and transcripts are kept in Prosper's own environment, or whether any de identification occurs before QA scoring. Safety engineering is partly described: an agent that is unsure escalates, and language suggesting an emergency stops the automation and directs the caller to emergency care.

Business associate status is stated and supported. Prosper says it signs business associate agreements as standard and publishes its HIPAA posture: SOC 2 Type II, encryption in transit and at rest, single sign on, a zero day retention agreement with OpenAI and a 30 day backup retention policy, and it tells buyers that compliance requires subprocessor agreements with every service touching PHI (model, speech to text, text to speech, telephony). The agreement itself is not published.

A recognized certification is named with its type. Prosper states SOC 2 Type II, alongside encryption in transit and at rest (AES 256 and TLS), single sign on, role based access and audit logging, and business associate agreements as standard. No trust portal publishes the report or its period.

No FDA pathway applies and none is claimed. Voice automation of scheduling, eligibility, authorization and billing calls sits well outside Software as a Medical Device.

What the company leaves unsaid is specific and consequential. This product runs artificial voice calls in both directions: to payers, which is business to business and outside the consumer telephone consent framework, and to patients, which is not. The FCC ruled in February 2024 that AI generated voices are artificial voices under the Telephone Consumer Protection Act, requiring prior express consent and identification of the entity responsible for the call. Outbound patient contact for balance collection and proactive re engagement sits squarely inside that.

Recording consent is a second exposure, since all party consent statutes in several states cover business calls as well as consumer ones, and this product generates recordings of both.

The silence stands out here, because the company publishes detailed guidance on health privacy compliance and subprocessor chains while saying nothing about the regime governing its own outbound dialing.

Multilingual operation is marketed with no subgroup performance behind it. Prosper's FAQ says its agents support up to 45 languages and understand a wide range of accents, dialects and speaking styles, and that elderly patients often prefer the AI; no recognition accuracy or resolution rate is published by language, accent or age. A caller the system fails to understand leaves no complaint and no record. No governance framework or bias evaluation is published.

Mechanisms exist and nothing stands behind them. Prosper says every call is reviewed by AI powered QA that scores accuracy and compliance, and passes call recordings and transcripts back to the customer's systems, so a customer can check what an agent did. It cites a 99.9 percent uptime SLA, which covers delivery, not the correctness of what the agent obtained or said. No accuracy figure with a method, warranty or remediation commitment for a wrong benefits or authorization result is published.

Named systems with read and write described, without documentation to verify. Prosper's FAQ says it integrates with athenahealth, eClinicalWorks, AdvancedMD and Epic through the integration pathways those systems support, and writes back appointment changes, patient and insurance details, benefits verification outcomes, staff notes, recordings and transcripts. No marketplace listing or integration documentation is published.

Options are stated and residency is not. Prosper supports cloud or on premises deployment and connects to a practice's existing phone system, and names OpenAI as its model provider with zero day retention. No hosting provider, region or tenancy is published for the cloud option, and the speech and telephony providers that carry calls are not named.

There is no public pricing. Prosper sells through enterprise agreements with provider groups and billing organizations and publishes no rate card.

Coverage is named with validation behind part of it: outpatient groups across named specialties (behavioral health, cardiology, dermatology, ENT, gastroenterology, OB/GYN, ophthalmology, orthopedics, pain management, pediatrics, podiatry, primary care, urology, multispecialty), billing organizations and health systems, covering scheduling, intake, benefits verification, prior authorization, claim follow up and billing calls. Anonymized results support OB/GYN and GI scheduling; payer call outcomes are not validated by specialty. No clinical claims are made.