Question 1

What is Praxigen?

Accepted Answer

Praxigen is an AI-native workspace that helps specialty practices handle prior authorization end to end, from looking up each payer's requirements, to checking documentation and codes before you submit, to drafting policy-cited appeals when a denial hits. It is built for the front-line PA team, and it works without a six-month EHR integration.

Question 2

Insurers are using AI to deny faster. How does provider-side AI actually help?

Accepted Answer

Payers have automated their side of prior auth, generating denials in seconds. Most practices are still fighting back with portals, faxes, phone trees, and institutional memory. Praxigen levels the field: it gives your team the same speed on the provider side, checking every request against the payer's own documented criteria before you submit so fewer get denied, and drafting evidence-backed appeals the moment one does. You are no longer out-gunned by an automated system.

Question 3

How do you prevent the AI from making things up (hallucinating)?

Accepted Answer

This is the core of how we built Praxigen. The AI does not answer from memory, it retrieves the real source first (the payer's policy, the governing CMS or clinical guideline, the specific clinical-trial evidence) and is required to cite it, and a verification step checks that the citation actually supports the claim. Every recommendation traces back to a real document you can click. If we do not have a trustworthy source, the system says 'verify against the source' rather than guessing.

Question 4

What data is Praxigen actually querying?

Accepted Answer

Public, authoritative sources, not a black box. That includes payer medical policies and coverage criteria, CMS coverage rules (LCDs and NCDs), and clinical evidence from ClinicalTrials.gov, mapped to the specific procedure and payer. We keep this current as payers and rules change. We do not scrape proprietary licensed criteria or use made-up data: if it is in Praxigen, it is a faithful summary of a real, cited source.

Question 5

Is Praxigen HIPAA compliant? How do you handle patient data?

Accepted Answer

Praxigen is not yet HIPAA-covered, and there is no BAA in place yet, so we are upfront: the pilot runs on de-identified or sample data only. Patient identifiers are stripped on your device, with a human-review step, before anything is sent, and you stay in control of what is submitted. When we move to handling identified records, we sign a BAA and put full HIPAA controls in place first. We would rather tell you exactly where we are than overstate it.

Question 6

Doesn't de-identifying the note lose the clinical detail the AI needs to be accurate?

Accepted Answer

It is a fair concern, and the honest answer depends on the threat model. The well-documented accuracy ceiling for de-identification comes from research and dataset releases, where many records are pooled and contextual quasi-identifiers can be re-linked to re-identify people. Praxigen's situation is different: it is one patient's record, de-identified on your device, for a single stateless model call. There is no dataset of other patients to cross-link against, so that ceiling does not bind the same way. Just as important, identity and clinical signal are not the same field. The on-device step removes the exact identifiers (name, MRN, date of birth, address) while preserving the clinical facts that decide the case: the patient's age, the trend over time, medication names, lab values, exam and imaging findings. We strip what identifies the patient, not what makes the medical-necessity argument. And the realistic alternative is not perfect de-identification, it is staff pasting fully identified records into consumer chatbots with no redaction at all. Removing identifiers on your device first is a strict improvement on that, while we build toward full HIPAA controls and a signed BAA for handling identified records.

Question 7

Does the AI submit to payers on its own?

Accepted Answer

No. Praxigen prepares the submission and the appeal; a person on your team reviews and submits. Every output is an AI-generated draft for human review, we never file anything under your providers’ names automatically. You keep full control.

Question 8

Do we have to integrate with our EHR (Epic, athenahealth, eClinicalWorks) to start?

Accepted Answer

No, and that is deliberate. There is no EHR integration, no Epic build, and no IT project required to pilot Praxigen. Your team pastes the clinical note or uploads it as a PDF, Praxigen de-identifies it on your device with a human-review step, and the tools work the same day. Nothing forces your staff to leave their charting system and double-document. Native EHR integration over FHIR is on the roadmap, athenahealth first, and we are honest that we are not yet listed in Epic’s app marketplace; deeper integration comes only when it genuinely removes steps and the compliance pieces (a BAA and full HIPAA controls) are in place.

Question 9

Can Praxigen read PDFs and outside records (PT notes, MRI reports, scanned charts)?

Accepted Answer

Yes. You can upload a PDF directly and Praxigen reads it, which matters most where an authorization hinges on documents that live outside the main chart: physical-therapy notes that show conservative care was tried and failed, outside MRI or radiology reports, and scanned referrals. The note checker scores that documentation against the payer criteria and flags what is missing before you submit. Everything is de-identified on your device first.

Question 10

How much does it cost?

Accepted Answer

The founding pilot is a 6-month design partnership: the first 2 months are free, then a flat $299/month founder rate for the whole practice (any size), locked for the term and at renewal. At month 2 you get a KPI report on your results and choose whether to continue; opt out and you owe nothing.

Question 11

How does the pilot program work?

Accepted Answer

You onboard with hands-on, white-glove setup and we capture your baseline. For the first 2 months you use the workspace free, with weekly check-ins. At month 2 we deliver a KPI report on the time saved and gaps caught; you decide whether to opt in. Months 2 through 6 run at the $299/month founder rate with continued support. De-identified or sample data throughout.

Question 12

Can we start with one provider or one procedure to keep it low risk?

Accepted Answer

Yes, and we recommend it. A good pilot starts narrow: pick one provider and one high-volume procedure type (for example total knee arthroplasty or lumbar fusion), run it for about 30 days, and look at whether it is actually saving your team time and catching gaps before you expand. You are a design partner, not a long-term contract, and at month 2 you get a KPI report on your own results and decide whether to continue. Our guide on running a low-risk PA pilot has the full checklist.

Question 13

What makes Praxigen different from other AI tools?

Accepted Answer

Three things. (1) It is source-grounded: every answer cites the real payer policy or clinical evidence it came from, so your staff can stand behind it. (2) It is end to end: lookup, documentation and code checks, tracking, and appeals in one place, not an appeals-only point tool. (3) It gets sharper over time: Praxigen maps which evidence actually satisfies which payer's criteria, so its guidance improves as more cases run. The data behind it is the hard part to copy, not the interface.

Question 14

Does Praxigen guarantee my prior auths get approved?

Accepted Answer

No, and we will not claim that. No one can guarantee a payer’s decision. What Praxigen does is help you submit complete, policy-aligned documentation the first time and build evidence-backed appeals, which is how avoidable denials go down. We help you win the cases that are winnable, and clearly flag the ones that are not.

Question 15

Who is Praxigen for?

Accepted Answer

Small and mid-sized specialty practices with high prior-auth volume: ortho, spine, pain, and similar. The day-to-day users are the people who live in PA: prior-auth specialists, medical assistants, billing and front-desk staff. The buyer is usually the practice or office manager.

Question 16

Which payers and procedures do you cover?

Accepted Answer

We currently cover 28 payers' documented requirements - including national plans and the major Massachusetts/Boston-market payers (Blue Cross Blue Shield of MA, Harvard Pilgrim, Tufts, Mass General Brigham Health Plan, WellSense, Fallon, MassHealth) - across orthopedics, spine, pain, physical therapy, podiatry, radiology, cardiology, and now rheumatology, dermatology, and gastroenterology biologics. We add coverage continuously. On a demo we will check your specific payers and procedures live.

Question 17

Is our data used to train AI models?

Accepted Answer

No. We do not sell your data or use your patient information to train third-party models, and during the pilot you are working with de-identified or sample data anyway. What improves over time is our mapping of evidence to payer criteria, built from public sources and de-identified, aggregate outcomes, never patient-level records.

Question 18

Who's behind Praxigen?

Accepted Answer

A team building the prior-auth tool practices kept asking for: student-founded out of the University of Michigan, 1st place at Build & Pitch (NYC Tech Week 2026), and working directly with front-line PA staff to shape what ships.

What is Praxigen?+

Insurers are using AI to deny faster. How does provider-side AI actually help?+

How do you prevent the AI from making things up (hallucinating)?+

What data is Praxigen actually querying?+

Is Praxigen HIPAA compliant? How do you handle patient data?+

Doesn't de-identifying the note lose the clinical detail the AI needs to be accurate?+

Does the AI submit to payers on its own?+

Do we have to integrate with our EHR (Epic, athenahealth, eClinicalWorks) to start?+

Can Praxigen read PDFs and outside records (PT notes, MRI reports, scanned charts)?+

How much does it cost?+

How does the pilot program work?+

Can we start with one provider or one procedure to keep it low risk?+

What makes Praxigen different from other AI tools?+

Does Praxigen guarantee my prior auths get approved?+

Who is Praxigen for?+

Which payers and procedures do you cover?+

Is our data used to train AI models?+

Who's behind Praxigen?+