In partnership with
Issue #14 · Ops & Om · Health of Business series · Episode 1
or listen on Spotify · Apple Podcasts
I want to tell you about the best hire I made this year.
It works three shifts a day. It never calls in sick. It costs me almost nothing. And its entire job is one thing: my email.
Let me describe an inbox and you tell me if it sounds familiar. Six receipts from six vendors that need filing by which business they belong to — because at tax time your accountant doesn’t care that it landed in your personal Gmail. Client threads you can’t drop. A partnership pitch that might be real and might be spam wearing a suit. Your spouse forwarding you the house thing that needed handling yesterday. A medical appointment buried under four newsletters you haven’t opened in a month. Now multiply it by three, because you have three email addresses. Maybe five.
The standard advice — “check email twice a day,” “get to inbox zero” — is just discipline advice.
Discipline doesn’t scale. Systems scale.
So I built a system. The first Ops & Om episode is the full walkthrough. Here’s the short version.
Here’s what we’ll cover in this issue:
The Real Project Isn’t “Get AI to Do My Email”
When I tried to build this before, I had the project wrong. I thought it was get AI to do my email. It’s not.
The project is: write down how I make email decisions — so clearly that someone else could make them for me. AI is just “someone else.”
The whole thing is three pieces: a rulebook (a literal text file that says what to do with every type of email), labels (Gmail labels that act like folders), and a scheduled task that runs at 8am, noon, and 4pm — it reads the rulebook, processes the inbox, and writes me a digest.
Week one, it asked about everything, like an intern shadowing me. By week three it handled about 90% silently and only brought me the actual decisions.
I didn’t train a model. I wrote an operations manual — one question at a time.
Try this prompt:
Help me build an email triage rulebook. Ask me one question at a
time about how I handle different kinds of email (receipts, client
threads, pitches, newsletters, personal). After each answer, write
the rule in plain language and add it to a running rulebook. Don't
move on until each rule is clear enough that someone else could
follow it without me.The Four Guardrails That Make It Safe
People ask: you let AI loose on your email? Your email has your whole life in it.
I do — because of four rules that are non-negotiable, written at the very top of the rulebook so it reads them every single run:
Never deletes. Archiving means “remove from my main inbox.” Nothing is ever destroyed.
Never sends. It drafts replies in my voice — and they’re good — but every one waits in my drafts until I hit send.
Never touches settings. No forwarding changes, no filters. Nothing.
Never guesses. Ambiguity goes into the digest as a question, not an action.
Run the worst case on that. An email stays in my inbox. That’s the catastrophic failure mode — one extra email.
You don’t trust AI by hoping it behaves. You design a box where misbehavior is impossible, then let it earn scope over time.
Try this prompt (paste these at the top of your rulebook):
NON-NEGOTIABLE RULES — read these every run:
1) Never delete. Archive = remove from inbox only. Destroy nothing.
2) Never send. Draft replies in my voice; leave them in Drafts.
3) Never change settings, filters, or forwarding.
4) Never guess. Put any ambiguity in the digest as a question.The Wins — and the Boring Fix for Too Many Inboxes
One week, three things I didn’t have to think about:
It filed an ad receipt to the right company’s label — a clean look at tax time instead of a weekend project.
We closed on a property. It filed the whole deal thread, flagged “transaction funded,” pulled the fine print I usually skim past (no lender escrow means nobody mails you the tax bill — check the county site in October), and set a calendar reminder for October.
A retail chain with 86 stores wanted to carry our products. It flagged the email as real revenue and drafted a reply in my voice with the two qualifying questions I’d have asked. I reviewed, sent, done.
And all those accounts? Solved with plain Gmail forwarding. Every address forwards a copy into one hub inbox. Originals stay untouched, nothing moves, nobody changes habits. One hub, one rulebook, one digest.
Email is the front door to your operations — which makes it the highest-leverage thing to systemize first.
Get the exact system — free
I packaged the whole thing into a free kit: the rulebook template with the guardrails baked in, the scheduled-task prompt word for word, and the setup checklist (including the multi-account forwarding piece).
Rather have it done for you? I build these for clients — single inbox or full multi-account with receipts auto-saved to your Drive. Book a 1:1.
Baldomero Garza — Find me on X, LinkedIn, Instagram, or book a 1:1.
P.S. — This is episode one. If you only do one thing this week, watch the walkthrough and start your rulebook — it takes an afternoon. Watch on YouTube →
Your prompts are leaving out 80% of what you're thinking.
When you type a prompt, you summarize. When you speak one, you explain. Wispr Flow captures your full reasoning — constraints, edge cases, examples, tone — and turns it into clean, structured text you paste into ChatGPT, Claude, or any AI tool. The difference shows up immediately. More context in, fewer follow-ups out.
89% of messages sent with zero edits. Used by teams at OpenAI, Vercel, and Clay. Try Wispr Flow free — works on Mac, Windows, and iPhone.