Grounding tier
Last updated: July 2026 (Beta)
Beta. DociCA is in invite-only beta. These terms apply to beta use. DociCA is document-automation software, not a law firm, and does not provide legal advice. See our Beta Disclaimer.
A grounding tier is a label carried by every AI review finding, recording how strongly that finding is evidenced.
Why it is not a confidence score
A confidence score reports how sure a model is. A grounding tier reports what was actually checked— a different thing, and a verifiable one. A finding can be stated confidently and still rest on nothing, so DociCA records the evidence rather than the model’s certainty.
The three tiers
Verified text. The finding quotes a span of government legislation that was confirmed to appear, literally, in the real source. This is the only tier DociCA treats as evidence-backed.
Verified citation. The statutory citation attached to the finding was verified, but the finding does not carry a verified quoted span.
Model generated. The default. The finding was produced by the model and nothing behind it was verified. Labelling this tier honestly, rather than hiding it, is the point: a reader can see exactly which findings rest on checked law and which do not.
How to read a review
Tiers are counted per review, so a report says how many of its findings are verified-text, verified-citation, and model-generated. Findings are information for your own review — DociCA is document-automation software, not a law firm, and this is not legal advice.