YMT Workbench
AI: …

Roadmap

What this prototype shows today, and where a production build goes next.

Download the one-pager (PDF)

In this demo

  • βœ“Application Check β€” cited compliance report with a deterministic rules engine + AI narration + validator pass
  • βœ“Lender Match β€” scored shortlist across the panel with cited reasons
  • βœ“Ask AI β€” adviser-to-adviser policy answers with citations
  • βœ“Fact Find Check β€” per-applicant document checklist + drafted client email
  • βœ“Policy library β€” public lender corpus, every claim traceable to source
  • βœ“AI usage & cost metering

Phase 1 β€” Pilot (4–6 weeks)

  • β—†Connect to YMT’s Trail: pull fact-finds & document packs straight in (no upload step)
  • β—†Load YMT’s full adviser policy library (aggregator access) behind the same citation engine
  • β—†Vision OCR for scanned documents and bank statements
  • β—†Per-adviser logins, with an audit trail of every check

Phase 2 β€” Automation (Pulse)

  • β—†Refix & rate-lock expiry reminders across the whole book
  • β—†Settlement & pre-approval-expiry chasers
  • β—†Document-collection nudges to clients when items are outstanding

Phase 3 β€” Inbox triage (Rachel)

  • β—†Read inbound client enquiries, draft a grounded reply, route to the right adviser
  • β—†Confidence-gated: auto-send only when certain, otherwise adviser approves/edits

Phase 4 β€” Compliance & scale

  • β—†FSLAA / CoFI advice-file export per application
  • β—†Multi-branch dashboards (QT, Wānaka, Invercargill, Nelson Tasman)
  • β—†Insurance & KiwiSaver advice surfaces for YMT Wealth