What it does
Agentic AI for health plans and providers covering prior authorization, utilization management, risk adjustment coding and fraud, waste and abuse detection. The fields below record what could be confirmed from the 3 public sources cited at the foot of this page; nothing is estimated to fill the gap.
Key facts
| Segment | Payer Prior Auth and UM AI |
| Website | |
| Not published | FoundedHQHeadcountOwnershipCertifications |
Medrecords AI vs basys.ai at a glance
Same 8 rubric rows, side by side. Vendor cells repeat what basys.ai publishes; a row it has not published stays that way. Medrecords AI cells are our own published terms.
| Rubric row | Medrecords AI | basys.ai |
| Processing model | AI drafts and cites; you review and sign. Every line cites its source page. | AI matches prior authorization requests against patient charts and policy criteria and returns decisions with an audit trail |
| Human QA | Yours, not ours. Output is editable and every claim links to the page it came from. | Not published |
| Typical turnaround | Minutes to hours per file | Claims 4x faster time to decision and risk adjustment chart review in 3 minutes against 30 manually |
| Pricing model | Pay per page, no subscription (Self-Service); annual licence for Enterprise On-Prem | Enterprise sales by demo request; no public pricing |
| Indicative cost | from 10¢ to as low as 5¢ a page, duplicates free | Not published |
| Export formats | DOCX, PDF and HTML, hyperlinked back to every source page; API or SFTP delivery | Not published |
| Best for | Law firms, IME clinics, legal nurse consultants and small TPAs | Health plans and payers running prior authorization and utilization review; providers also served |
| Customer satisfaction | We publish no rating of our own | Not published |
The boundary on our side does not move: the software drafts and cites; you review, you revise, you sign.
Evaluation criteria
Profile notes
What to confirm before you sign
basys.ai publishes 4 of the 8 fields in the Content Hub rubric. Across all 270 vendors profiled in the Content Hub, the median is 5 of 8.
The other 9 questions, and what all 270 profiled vendors do and do not publish, are in the evaluation guide.
