Where this stands
Case study in progress. Target completion: 2026-11-15. The full writeup lands once the research, build, and eval phases are done.
The customer scenario
Financial services or legal-tech customer building consumer-facing AI in a high-stakes regulated domain (estate, tax, healthcare, insurance) needs structured tool-calling, document understanding, compliance, and trust-by-design UX.
My product canvas
This case study is grounded in LastingPath (live at lastingpath.com) — an estate administration platform that replaces $300/hr attorneys with a structured workflow. Sage (Claude Sonnet reasoning layer) is the brain — not a chatbot, a router that analyzes estate specifics and recommends applicable tasks. Canonical writeup: kaydenlabs.com/work/lastingpath.
Architectural patterns to be demonstrated
- Claude Sonnet reasoning layer (Sage) as a structured workflow router — not open-ended generation; conditional routing on jurisdiction, asset type, filing requirements
- OCR-driven structured form auto-population — Tesseract.js for death certificates and financial docs, feeding 47+ guided wizards (IRS Form 56, Form 706, RUFADAA-compliant social media packets)
- Trust-by-design UX with full legal-event audit logging — every decision logged, AES-256-GCM field-level encryption for sensitive fields
- Cron-based deadline sequencing for time-bound legal obligations — creditor claims, tax filings, cease-and-desist windows
What you'll find here when it ships
- Architecture diagram
- Eval set with results
- Cost and latency analysis
- Failure modes documented
- "What I'd do differently" retro