How verdicts work

Every verdict comes from your rules in Criteria. AI reads the solicitation and pulls out the facts: codes, set-aside, value, deadline, place of performance and requirements. It also judges whether the bid fits your lanes (what you sell or do). When the source publishes a fact itself (for example SAM.gov's set-aside or deadline), Greenlight uses the source's value.

Those facts run through a fixed checklist built from your Criteria:

• NO-GO: any rule set to Auto NO-GO fails, or the response deadline has already passed.
• REVIEW: a rule set to Flag for review fails, or a fact your rules need wasn't stated in the document. A missing fact never gives GO.
• GO: everything on the checklist passes.

The lane judgment can change the verdict. A bid outside your lanes follows your "Opportunities outside your lanes" setting: Auto NO-GO or Flag for review. When the lane read is uncertain, Greenlight leans to REVIEW ("Borderline fit"), never GO. A "Not ..." exclusion in your lanes always wins over a match.

The same facts and the same rules always give the same verdict.

AI notes appear separately and are labelled "AI note". They never change the verdict. If the AI spotted something specific, such as a large performance bond, you may see it next to the verdict, for example "GO · 1 flagged risk".

Next to every verdict, Greenlight shows whether the three key facts (value, place of performance, deadline) were stated: "All key facts stated", "1 key fact missing" or "Low data".

A verdict is a screening aid, not a bid decision. Always read the solicitation before you bid.