OpenAdapt's only public audit-trail example is labeled fictional data

OpenAdapt, an open-source desktop automation agent used in at least one healthcare billing-audit deployment, describes its core governance claim this way: "OpenAdapt turns a demonstrated task into an inspectable program. It checks the agreed result and stops when verification fails," according to the company's site, reviewed by AI Dispatch Oct. 4. The only public walkthrough of what that verification looks like in practice, however, is explicitly labeled a work of fiction.

The site's product demo shows a six-step sequence — "Read approved input, Find the right field, Check record identity, Apply the change, Verify saved result, Check the actual result" — ending with the line "Read the saved record separately to confirm the change." Below the demo, the page states: "Interactive illustration with fictional data. No live system is connected." Nothing elsewhere on the site describes what a real, saved audit record from a live OpenAdapt deployment contains, how long it is kept, or whether it can be altered after the fact.

The company does disclose a real use case: it says "OpenAdapt collects evidence from Cerner PowerChart and reconciles it against report spreadsheets" for a physician's monthly billing audit, producing "a workbook of missing credits, with source evidence for review." Human sign-off is built into that workflow — "the physician reviews the findings before the billing follow-up," the site states — and the company says it will "keep each client's access and review rules in place" when operating across client systems.

OpenAdapt's underlying automation engine is open source: "the local OpenAdapt engine is MIT-licensed," the site states. Pricing for a deployment is negotiated per engagement — "we agree the scope, acceptance criteria, and price before the build. Ongoing maintenance is a separate agreement," according to the site.

AI Dispatch found no public evidence of a retained audit trail independent of the fictional demo and scored the audit-trail dimension of its trust-gap rubric accordingly. The rubric, which scores vendors zero to three across five dimensions with evidence required for every score, is published at aidispatch.news/methodology.