See it against your own claims.
The fastest way to judge Orukai is to point it at a de-identified extract of your own book and look at what it surfaces. Reach out to arrange a walkthrough.
Orukai reviews a health plan's claims prepayment, ranks what deserves a human, and shows its work. Every score is a sum of named rules with the reason each one fired — so an examiner can act on it, and an auditor can check it.
Four steps, each one recorded. Nothing happens to a plan's data that the plan cannot see afterwards.
Send claims and roster files in whatever columns your systems already produce. Orukai proposes a mapping; you confirm it. Nothing loads until you do.
Peer-billing outliers, duplicate submissions, off-roster providers, impossible dates. Thresholds are derived from your own book, not borrowed from someone else's.
Examiners see the claims that earned attention, ranked, with the evidence attached and the prompt-pay clock in view. Decisions are one click and on the record.
Provider-level risk catches what no single claim shows — the practice inflating one visit in five. Build the case with its full audit trail behind it.
Most risk scores are a number with a story attached afterwards. Orukai's are the opposite: points are earned by a rule that fired on data the plan actually sent, and the score is their sum. Nothing is scaled to hit a target. If nothing fires, the score is zero — and that is the correct answer, not a failure.
Illustrative claim. Layout and reason text as the product renders them.
Prepayment review, provider risk, investigations and the reporting that has to hold up afterwards.
Claims ranked by earned risk, filterable by line of business and status, with the prompt-pay deadline visible before it becomes a problem.
Patterns across a provider's whole book — the ones no single claim reveals. A practice that bills normally four times and inflates the fifth does not move a median; it moves this.
Promote a claim or a provider into a case, gather evidence with its citations intact, and track exposure and recovery through to resolution.
Coverage reporting that states plainly which rules fire, which are blind for lack of data, and what the model is not measuring. No metric without a computation behind it.
Your column names, your mapping, confirmed by a person. Rows that fail validation are quarantined with a reason and stored — never silently dropped.
Examiner, medical director, investigator and leadership each see the view their decision needs, over one set of numbers that reconcile.
Claim files carry members, providers and money. These are engineering decisions, not policy statements.
Every plan's data is separated by row-level security in PostgreSQL, applied to every query including those run by the database owner — not by a filter the application is trusted to remember.
Orukai reads what you send and writes nothing back. The ingestion layer has no ability to delete or move a file: your record of what you sent stays yours, intact.
Every approve, deny, pend and rescore is written to an audit ledger against the authenticated user who made it — taken from the session, never from anything the browser claims.
Diagnostics and failure messages are scrubbed of member identifiers, filenames and row content before they reach a log, because an error message is where claim data usually escapes.
The fastest way to judge Orukai is to point it at a de-identified extract of your own book and look at what it surfaces. Reach out to arrange a walkthrough.