The verification gate
The single point at which AI output becomes part of your record of truth. Every authoritative obligation in the system has passed through here.
Why it exists
clmSpace separates what the AI proposed from what the business stands behind, with one explicit step. Reviewers see the proposed obligation, the verbatim citation and the source document side by side in a single view, then verify, override or reject.
The Review Queue
Every AI-proposed obligation lands in the Review queue, which carries a live count in the workspace sidebar so reviewers always know what is waiting.
- Verify: accepts the AI value as it stands and stamps the verifier and the time.
- Override: corrects the value (for example, the AI read “Net 30” where the actual term is “Net 45”). The corrected value takes effect, and the original AI value and the reason are kept in the audit trail.
- Reject: removes an obligation that is not supported by the contract. It is withdrawn from the record while the citation is preserved for audit.
What changes once verified
- The obligation counts in dashboards, deviation detection, renewal packs and counterparty risk.
- It feeds the history signal that informs your contract standards.
- It can be cited by drafting flows and Ask clmAI.
Standards verification
The same model applies to the playbook itself. An AI-derived standard position (from a template or from your verified-obligation history) is held as a proposal until an admin confirms it on the Standards page, at which point it becomes an authoritative standard. See Tier model.