Area: mobieusKnow (audit p6) · Surface: /know/{slug} · Dimension: improve-existing · Severity: major
Article voting ('Was this helpful? Yes/No', or thumbs up/down) is a standard KB feature in Zendesk, Help Scout, Intercom, and Confluence. It does double duty: it gives readers a low-friction way to signal value, and it gives moderators a content-quality signal that drives the 'which articles need work' backlog. mobieusKnow has an AI stale-detector for staff (AdminAIController::wikiStale) but no reader signal at all, so the team is guessing about which pages actually serve members. Adding per-article up/down feedback (optionally with a free-text 'what was missing?' on a downvote) would feed both a public 'most helpful' sort and the existing AI gap tooling.
Evidence
There is no feedback capture anywhere in the Know subsystem. The schema has no knowledge_feedback / page_rating table (only knowledge_pages, knowledge_revisions, knowledge_user_trust, knowledge_moderators, knowledge_audit_log — schema.sql:3387-3487). The article template show.php renders contributors, reading time, word count, revisions, and related pages (show.php:159-285) but no helpful/not-helpful control. A grep for 'helpful' across the Know controllers/templates returns only forum/report credit code, nothing KB-related.
Suggested fix. Add a knowledge_feedback table (page_id, user_id nullable, vote enum up/down, comment nullable, created_at) and a 'Was this helpful?' control at the foot of show.php. Aggregate counts into a helpfulness score; expose 'most helpful' as a sort on the index and feed downvote comments into the stale/gap AI flow.
Filed by the automated tenant-app audit and adversarially evidence-verified. Status: verified. Open — not yet actioned.
Patrick Bass
@mobieus