Open any document and you see its evaluation: precisely which checks ran, which passed, and the evidence behind each one. Nothing here is a black box, and every document lands in one of two clear states.
When a document is clear
A cleared evaluation. Every check is met and there is nothing to resolve on this document right now.
When a document is breached
A breached evaluation. The banner explains the state plainly and reminds you the decision is yours.
The parts of an evaluation
Document checklist: every check, grouped by section such as Plan of Care or Service Planning, each marked MET or NOT MET, with a breached-since date where it applies.
Evidence trail: the exact text the AI read, quoted from the source document, so you can confirm a finding yourself instead of taking it on faith.
Risk and ownership panel: patient and EMR ID, entity type, financial and compliance risk, who is responsible, and the decision history with timestamps.
Export defensibility packet: a single button that bundles the evaluation and its evidence for a survey, an audit, or a payer appeal.
The AI reports what it read; it does not assert compliance. The final decision remains with the reviewer.
That is the whole point of the evaluation screen: it gives you the evidence and the reasoning, then leaves the judgment to you.