Intent Governance: When Bot-Like Intent Patterns Get Throttled

Intent matching only works if intents are real. A buyer who creates 200 intents and converts none of them is gaming the system — saturating sellers’ match feeds and not closing deals. Governance scoring is the throttle that prevents the worst patterns without blocking legitimate bulk-buyers.

What gets scored

For each buyer (or buyer-tenant), the platform computes a governance score from deal-history: ratio of intents to closed deals, ratio of opened deal rooms to acceptances, average time-to-first-action on a match, and intent volume per period. Low scores indicate spam-like behaviour. High scores indicate engaged buyers who close.

Tenant thresholds and enforcement

Tenant settings (/settings/market/governance) define the thresholds. When a buyer falls below the threshold, the F2b enforcement kicks in: their new intents land in a review queue rather than auto-matching live. The buyer is notified that their pattern triggered review. Sellers don’t see noise from buyers who don’t convert.

Admin override

Sometimes the throttle catches a legitimate bulk-buyer — a procurement team that creates many specific intents because their org has many specific needs. The admin governance page (/market/governance) shows the queue with each entry’s context, and the platform admin can override (release the intents into the live match feed) with a reason recorded for audit.

Why this isn’t a permanent ban

Scores recompute as deal-history accumulates. A buyer who closes ten deals after a slow start moves up the score curve and exits the throttle automatically. Governance is a friction layer, not a banning mechanism. The point is to keep the match feed useful, not to police buyers in perpetuity.