QuickIntell

AI agents are the engine of QuickIntell's core modules: payer portal bots that sign in and work eligibility, claim status and authorizations, QuickVoice calls to payers and patients, AI coding and the AI scribe, and the Agent Builder that lets operators create their own agents. Most of the platform's selling points rest on that automation.

The platform carries value beyond the models. It includes workflow orchestration, an Availity clearinghouse connection, a REST API, an EHR of its own, a large set of free coding and coverage reference tools built from CMS files, and an optional managed service in which people run the work. The agents drive the product, and the platform around them would still route and track revenue cycle work without them.

Write actions and high impact automations configured by the customer pass through approval gates before any change is committed, and QuickVoice contact workflows can be set to run automatically, semi automatically or only with approval, according to the customer's policy. Execution metadata is logged without patient values, and the pitch is built on keeping approvals and exceptions with the customer's team.

The thresholds themselves are left to each customer: nothing describes the default gates, which actions count as high impact, or what the agents may do without approval out of the box. How an agent decides it cannot finish a task and hands it to a person is also not described.

The architecture is described by function: bots that operate payer portals, voice agents, an AI assistant, coding and scribe models, workflow orchestration and an Agent Builder, connected through a REST API with an OpenAPI reference and a setup guide for AI clients over the Model Context Protocol. Agent memory is scoped to each organization, and recall over old messages is switched off so patient details cannot resurface by inference.

No language model, speech engine, provider or version is named, and nothing explains how models are updated or how customers are told of a change. The open API documentation serves technical buyers well, and it describes the platform's interfaces rather than what produces the AI output.

Cloud hosts are named, AWS, Microsoft Azure and Google Cloud in US regions, along with Stripe for billing, and the trust center lists a subprocessor document and an AI data handling document among the materials a customer can request. No language model or speech provider is named.

The stated terms help: no patient data is used for model training, and voice recordings and assistant threads stay under the US residency controls. Which model companies receive patient data during a run, and on what retention terms, is set out only in the documents released on request.

Case studies report outcomes for unnamed customers. A six surgeon orthopedic practice with three physician assistants is described as cutting no shows from 17 percent to 9 percent and adding $420,000 a year through pre authorization checks before imaging and surgery and better documentation, with no dates, baseline period or method behind the figures.

Customers in the case studies go unnamed, no third party evaluation is cited, and nothing measures the accuracy of the agents themselves: how often a bot's eligibility check or authorization submission is right, or how many runs fall out to a person.

QuickIntell states a no PHI model training policy and gives specific retention controls: clients configure record retention periods, unpinned AI Assistant threads are purged after 90 days, and audit evidence is kept for seven years where required. Voice recordings, transcripts and assistant threads are encrypted and scoped to each tenant.

De identification is not described, and nothing says how long agents' working data, such as payer portal screens captured during a run, is kept. Safety engineering is covered through approval gates on write actions and high impact automations, with no description of how a wrong submission is detected once it has been approved.

A business associate agreement is included on every plan, including the free trial, with no upcharge, and QuickIntell will red line custom agreements or sign the customer's own paper for health system and enterprise contracts. The trust center describes HIPAA aligned safeguards, a request path for the agreement and HIPAA privacy controls handled through contract and retention workflows.

The standard agreement is not posted to read before signing, so its terms, including how it binds the cloud and model subprocessors listed in the trust center's documentation, are seen only after a request. The trust center's subprocessor disclosure is likewise available as a document on request.

QuickIntell names a SOC 2 Type II attestation and ISO 27001:2022 certification, with the SOC 2 cover page under NDA and the ISO certificate available on request, and its pricing page also refers to HITRUST CSF disclosures. An annual third party penetration test is summarized with remediation tracking, and enterprise customers can request the full reports.

The trust center lists an incident response plan with drills and post incident review, role based access, tenant isolation and audit logging. Because the reports are released on request rather than opened from the page, dates and scope are confirmed only after asking.

QuickIntell sells revenue cycle software and makes no device claim. Its modules work under payer and federal billing rules, with free tools built on CMS edits, fee schedules, coverage determinations and the CMS hierarchical condition category model, and its coding, risk adjustment and documentation improvement modules apply those rules rather than make clinical decisions.

The AI scribe documents a clinician's visit for review, and the voice agents contact patients for administrative reasons. Nothing markets the platform as diagnosing or recommending treatment, so the administrative positioning matches what it sells.

Governance is described through policy and controls: board level security ownership, a no PHI training policy, approval gated write actions and an AI data handling document available in the trust center. No evaluation of agent accuracy, coding accuracy or scribe quality is shared.

For revenue cycle agents, uneven performance tends to show across payers, specialties and claim types rather than patient groups, and no breakdown of that kind is given. A buyer cannot see whether automation rates or error rates differ between commercial payers, Medicare and Medicaid plans, or between simple and complex claims.

Approval gates before write actions and execution logs for each run give a customer a way to stop and trace what the agents do. Overage alerts and plan flags protect the customer's budget rather than the accuracy of the work.

No guarantee, service credit or remediation commitment covers an agent's output, such as a denied claim or a missed authorization, and an incident service level is mentioned without its terms. Customers who want people accountable for results can buy the optional managed service, whose performance commitments are not described either.

Connections to payers are specific: Availity's API for eligibility, claim status, remittances and attachments, and bots for Availity, UnitedHealthcare, Medicare, Medicaid and some 300 payers. On the record side, QuickIntell says it connects with EHR and billing systems and offers its own EHR, QuickEHR, without naming the outside record systems it reads from or writes to.

No EHR vendor, marketplace listing or interface standard such as FHIR is named for those integrations. A practice cannot confirm from what is offered that its own record system is connected, or how charge and documentation data move into QuickRCM, until it asks for the integration map for its EHR.

Residency is stated precisely: protected health information for US customers stays on US infrastructure, in the US East and West regions of AWS and Microsoft Azure and the US region of Google Cloud. QuickRCM records, QuickVoice recordings and transcripts and AI Assistant threads are encrypted and scoped to each tenant under those controls, and Stripe handles billing data only, with no patient information.

Isolation is enforced at the data layer: every record carries an organization identifier and queries filter on it, so one customer cannot reach another's data. The service is hosted software with configurable residency options, and managed revenue cycle work is offered on top as a separate option. Encryption keys are held in AWS KMS and Azure Key Vault, databases use transparent data encryption, and residency for non patient business data follows tenant and agreement controls.

Plans are priced on the page: Solo Practice at $499 a month and Group Practice at $1,499 a month, with Health System and Enterprise plans quoted. Each plan includes a set encounter volume, overage is billed at a per encounter rate stated on the order form, and QuickIntell flags two months of overage in a row so a practice can upsize or stay on usage, with soft and hard billing alerts.

Add ons such as QuickVoice, the Agent Builder, specialty packs, credentialing and implementation services are line items on the same invoice. A free trial is offered, discounts apply to FQHCs, rural health clinics, behavioral health nonprofits, tribal health and public entities, and small migrations are included while larger ones are a fixed fee engagement.

The stated market is wide: solo and group practices, hospitals and health systems, revenue cycle management companies, and payers, MSOs, ACOs and labs, with specialty workflows searchable by specialty. Plans are built for each tier, from solo practices to enterprises, and modules span inpatient billing as well as practice workflows.

The outcome story offered concerns an unnamed orthopedic practice, and no hospital, health system, RCM company or payer is named as a customer. Breadth is claimed across every setting the product lists, with nothing yet validating most of them.