Vigil combines written review criteria, source-linked AI suggestions and versioned human decisions. Coverage and sample results stay visible alongside the review queue.
Work through case material under a shared, versioned review protocol.
Inspect cited passages, human decisions and validation samples before relying on the work.
Keep review decisions and an internal register with the case.
The receipt beside each step is an illustrative example of what a reviewer sees, not data from a real case.
Write criteria for responsive material and specialized review. Saving a change creates a separate protocol version rather than rewriting the instructions for an existing run.
criteria · protocol version · saved runTessera processes eligible source units and records analyzed, failed and withheld material. Missing units or failed sections keep the run incomplete.
analyzed · failed · withheldRetained AI claims carry quotes validated against the cited lines. A separate verification call checks the anchored passage; same-model checks are labeled as such.
verbatim quote · cited lines · check typeReviewers record responsiveness, privilege and redaction-review labels, tags and notes. Decisions keep their revision and actor. Prioritization reorders remaining work; it does not replace human review.
decision · reviewer · revisionInspect validation and elusion samples with their denominators, unknown labels and uncertainty. Sample metrics are not a certification of corpus-wide recall.
sample · denominator · uncertaintyDownload the review register in Word, PDF, Excel or CSV after recent authentication. Exports are audited and retain source references alongside human decisions.
source reference · human decision · auditSee how Vigil fits your case review. Sign in to your firm's workspace or request a walkthrough.