Civil-society outreach

The hub asks the people who already know.

Civil society won't learn another login. So the hub inverts the model: it detects where its knowledge needs verifying, emails a specific question to the experts best placed to answer (caseworkers, lawyers, hotline volunteers), and folds their replies back into prioritized context and new grading dimensions. This page shows that loop running live.

01

The loop

1CollectExperts opt in by email; only sha256 + topics persist.
→
2Detect gapsKnowledge that needs field verification, prioritized.
→
3SolicitA targeted question drafted for matching contacts; a curator dispatches it.
→
4ObserveReplies vetted (PII + intent), turned into context signals.
→
5PrioritizeCorroborated gaps rank up + propose new rubric dimensions.

Loading live outreach state…

02

Context gaps the hub wants verified

Each is a concrete, field-answerable question. Priority rises as observations corroborate it. Draft campaign previews the solicitation email and the matching opted-in contacts (it only sends when SMTP is configured on the deployment).

Loading gaps…

03

Prioritized context from observations

Ranked by how strongly civil-society replies corroborate each gap. Each carries a candidate grading dimension, so outreach feeds the evaluation rubric, not just the document corpus.

No observations yet. As replies arrive, prioritized context appears here.

04

Join the expert list

Opt in to receive occasional, specific knowledge questions. Your raw email is never stored on the hub — only a sha256 hash + your topics, for curator outreach planning. Unsubscribe any time.