workflows active: 44 last morning round: Thu 07:00 last deploy: 2026-09-03
Forbes / Operations WORK / 03

The operating playbook and AI

Lifecycle standards, prompt libraries, QA checklists, and governance: the operating system that lets a two-person team run 10+ concurrent programs, with custom GPTs doing the trainable work.

Problem

Ten-plus lifecycle programs, two people. Anything that lived in someone's head became a bottleneck. AI could take a lot of the repeatable work, but not without guardrails, because it'll produce confident garbage at scale if you let it.

What I built

Standards and QA gates

Standards first, on paper: what a program is, what every email has to contain, what counts as launch-ready. QA checklists gate everything before it ships, human or machine made.

An email template proof with the Forbes banner and a merge tag greeting, followed by quarterly planning notes
A template proof from the planning motion, merge tags and all.

Prompt libraries, custom GPTs, and governance

Prompt libraries for the recurring work, versioned like anything else we ship. Governance that says what AI is allowed to touch, what needs human review, and who owns the output. Custom ChatGPT Enterprise GPTs trained on those standards, and the trainable work runs through them now.

The operating playbook what runs on what 10+ concurrent programs run by a team of two Custom GPTs the trainable work runs through them; specialists keep the judgment calls Standards what a program is, what every email has to contain, what counts as launch-ready Prompt libraries the recurring work, versioned like anything else we ship QA checklists gate everything before it ships, human or machine made Governance what AI may touch, what needs human review, who owns the output yellow outline: the layer the trainable work runs through
The playbook, and what runs through it.

Two buckets and value reporting

A two-bucket contractor model: trainable execution in one bucket, real expertise in the other, and outside rates only get paid for the second. And value reporting for the martech stack, so every tool has to show what it contributed.

The two-bucket model operating playbook / who does what One bucket contractor work, undifferentiated Two buckets trainable execution: custom gpts, one operator real expertise: outside rates before after
One bucket, then two.

Stack

chatgpt enterprise (custom gpts) / claude code / n8n / smartsheet / monday.com

Outcome

Campaign work modeled at 149 to 206 specialist hours has been delivered in roughly 71 to 74 hours by one AI-assisted operator: about a 60% reduction, projected to reach about 79% once the lifecycle agents launch. The modeling is against specialist-hour benchmarks, and everything still passes the same QA gate before it ships.

Next: Stedd, the founder page