Case Study · 03 · Human-gated automation, in production
One real business, rebuilt AI-native · The Peaceful Eating Way · this is its operations arm

Sixty-six owed sessions. Then zero, on autopilot.

Clients of a coaching practice each expect their session recording. Manual delivery had quietly fallen sixty-six emails behind. The fix wasn't more effort — it was an agentic pipeline with a human at the one gate that matters.

66 → 0
Backlog cleared — and held at zero
4
Systems orchestrated daily
0
Wrong sends, ever

The problem

Deeply personal recordings. Zero margin for error.

These aren't meeting notes. They're recordings of people doing vulnerable inner work — and a recording sent to the wrong person isn't a typo. It's a privacy breach.

Sessions landed across two separate video-call accounts and a phone line with no API at all. Recordings were named inconsistently — sometimes a first name, sometimes a room name, sometimes a caller-ID that was simply wrong. Each one had to be matched to the right client, its link checked, and a personal email written and sent. Done by hand, it consumed hours a week — and the backlog had grown to sixty-six clients waiting.

The obvious automation — match and blast — was never on the table. The whole design question was: how much can the machine do while a human still approves every single send?

The build

The machine does the work. A human owns the send.

A daily pipeline that harvests, matches, validates, and drafts — then stops, and asks.

Zoom · 1:1s Zoom · groups Phone system no API — parsed from email Identity match calendars + alias table Validate links standing rule set Draft emails personalized, per client HUMAN APPROVAL nothing sends without a named, approved list Queued sender releases approved drafts Read-back every send A do-not-send exceptions list is enforced upstream of everything — some clients are never automated, by standing rule. The daily output is one review email to the owner: proposed sends, flagged links, unmatched recordings. Approval happens in plain language.
Four systems, one daily rhythm. The gold gate is the design: the machine prepares everything; a human authorizes every client-facing send.
The governance

Every rule was earned by a real incident.

Review-first isn't a default that shipped with the software. It's a ruling — and the validation rules under it each exist because reality demonstrated the failure once.

I.

No client email sends without explicit human approval.

Auto-send stays off until trust is earned. The pipeline proposes; the owner disposes — in plain language, from one daily review email.

II.

Every link is validated from the primary source.

A truncated link once reached a client. Now links are extracted only from full source content, checked for access barriers that would lock a client out, and paired with per-source download instructions.

III.

Identity is matched, never assumed.

Caller-ID names can simply be wrong. Matching runs against calendar records with a maintained alias-correction table, and anything ambiguous is surfaced for a human ruling instead of guessed.

IV.

Send is verified by read-back.

A success response proves the call succeeded — not that it did what you meant. Sent mail is confirmed by checking what actually left, not what was requested.

The proof

The failure hit. No client felt it.

Production incident · absorbed by design

The scheduler went silent for nine days.

The daily review job simply stopped producing its email — the quietest kind of failure, the kind that ruins automated systems. Because every send is human-gated, silence produced nothing wrong: no misdirected recording, no stale link, no client contacted in error. The gap was detected, the nine days were swept in a single validated batch — every link checked, every identity confirmed, approval given in one message — and clients were caught up. The failure was absorbed, not amplified.

That's the test that matters for agentic systems. Anyone can demo the happy path. The design question is what happens when a component fails silently on a Tuesday — and the answer here was: nothing a client would ever notice.

Why this matters to a venture builder
Agentic automation × a human at the gate × safe failure = trust that survives production

This is the architecture every agentic-AI company is trying to figure out: autonomy where it's safe, human judgment exactly where stakes are real.

Bring me the messy problem