♥
Prospective foster carers
Conversational onboarding in your own words — no 200-field forms. Your narrative is structured by AI, then reviewed and approved by your assessing social worker.
Begin onboarding →Foster Match helps UK Local Authorities structure foster carer and child narratives into auditable profiles, then ranks compatibility using peer-reviewed research — while keeping statutory placement authority with caseworkers.
Foster Match sits between recruitment conversation and placement shortlist — making Form F Section B semantically searchable for the first time in most LA stacks.
Conversational onboarding in your own words — no 200-field forms. Your narrative is structured by AI, then reviewed and approved by your assessing social worker.
Begin onboarding →Ingest referrals, review extracted profiles in a structured queue, and (phase 2) request ranked shortlists with per-dimension breakdown and override controls.
Open caseworker portal →A four-stage pipeline with mandatory human review at every decision point.
Carer interview or caseworker document upload. Raw text stored with audit trail.
LLM maps narrative to Form F–aligned JSON schema with confidence notes for uncertain fields.
Assessing SW approves or corrects extracted fields before any profile enters matching.
Hard legal gates → weighted needs score → semantic narrative similarity (phase 2).
Placement stability research consistently points beyond bedroom counts.
Foster Match encodes what the literature says matters — with transparent weights
and explicit not_a_placement_decision outputs.
Live today: narrative onboarding, AI extraction with confidence notes, caseworker review queue, and pgvector embeddings for future semantic matching.