[01] · Village Hale Technologies Inc. · 2026 — ongoing
Village Hale — a number your family texts
My current focus. A messaging-first family chief of staff — parents text one number and Hale takes over the invisible admin: it watches city registration windows across 15 GTA municipalities (the 7 a.m. openings that fill before breakfast), plans the week, and graduates from suggesting to executing as it earns trust, action by action — an independent reviewer agent gates every act, and everything leaves an immutable receipt. No app required; the dashboard is just the receipts room. Privacy is the product — PIPEDA and Quebec Law 25 compliant by default, teen data redacted by construction. Live on a real number since August 2026. Empty repo to production, solo.
The problem
Families lose their village when they move — the neighbours and elders who knew which class was worth it, who said this hard week was normal. That trust still exists in word-of-mouth between parents; it just doesn't scale and you can't reach it. The AI is the easy part. The hard part is rebuilding that trust online while handling the most sensitive data there is — a newborn's — without ever leaking it.
[01]
Architecture
Passive, event-driven, multi-agent — the app is the easy 20%.
Hale isn’t a chatbot you open. It’s a passive system: the web app enqueues events, a worker drains them off a Postgres-backed queue, and a pipeline of specialised agents classify, infer, draft, and review — each testable on its own, none trusted on prose alone.
The LLM is maybe a fifth of the system. The rest is orchestration, guarded tool interfaces, audit logging, and evals — the parts that decide whether an agentic product is trustworthy or a demo.
- web ↔ worker split across one typed event contract
- the reviewer agent must invoke verification tools before it can approve
- every action writes an immutable audit row
[02]
The hard rule
Privacy by construction, not by policy.
This handles newborn and child data, so the default is the most restrictive thing that still works. Discovery sees only a coarse area — a postal prefix, never an address. A teen’s raw content is redacted from parents by default; only a category surfaces, and raw access is a logged, time-limited grant.
It’s enforced in code, not on a privacy page: a render-strip test fails the build if a child’s name could reach a session recording, and redaction sits at context-assembly so it can’t be forgotten downstream. PIPEDA, Quebec Law 25, Canadian data residency — the compliance core, not the fine print.
[03]
The moat
The village is the point; the AI just runs it.
Every other product leads with the assistant. Hale leads with the village — the genuinely good local things families near you actually do, vouched for by parents like you, more useful with every family that joins. The concierge finds and organises it; the network is what compounds.
Recommendations are grounded and cited — real venues, real sources — never a confident hallucination about where to take a toddler.
[04]
Operating principle
Autonomy is earned, never assumed.
New families start at observe-only. Hale drafts; the parent approves every action. Only after five clean approvals of one kind of task may it handle that kind on its own — and any automation can be withdrawn instantly. It never spends past a cap, or acts across both parents’ data without consent.
A household assistant that acts without permission isn’t a feature — it’s a liability. Trust is the product.
[05]
Discipline
Model tiers chosen by eval, not by vibe.
Every agent role has an eval — real cached responses, no mocked LLMs — that scores it across models. When “just use the biggest model” was tempting, the matrix showed Sonnet matched Opus on every role, and that Haiku silently failed the teen-safety classifier. So the tiering is measured: the cheapest model that holds the bar, the strongest only where it provably helps.
The takeaway
“The model is 20% of the system. The other 80% — privacy by construction, earned autonomy, grounded recommendations, eval-gated everything — is what makes an agent for a family's most sensitive data something you'd actually trust.”
Next case study
TripFix — flight-claim co-pilot →