This website uses cookies

Read our Privacy policy and Terms of use for more information.

In partnership with

Issue #23 · Ops & Om · Health of Business

You said it out loud, and that is the part that counts to the person who heard it. Somewhere in the last month you told someone you would send them a link, a study, a name. You meant it. Go check whether it went.

I run podcast pipelines for practitioners, and every host does the same thing on air. They get excited about a study or a protocol or a tool, and they say "I'll put that in the show notes." By the time anyone sits down to edit, four of those have gone by. Nobody remembers three of them.

I did it too, for years, before I built the thing that catches it. The promise felt small when I made it. It was small. Small promises are the ones that compound, because nobody makes a fuss about a single one and everybody quietly keeps count.

Here's what we'll cover in this issue.

The gap is small, which is exactly why it costs you

Nobody fires you over a missing link. That is what makes this so hard to see. There's no complaint, no email, no moment where someone tells you what changed. There's just a slightly shorter reply next time.

Here's the mechanic. When you promise something out loud, you create an expectation with a deadline the other person invented. They think "this week." You think "when I get to it." On day nine they stop expecting it. On day ten they've updated their model of you, and the update has nothing to do with the link. It's about whether the things you say map to the things you do.

Practitioners take this harder than most, because the whole business runs on being the person who is careful. You are careful in the room. Careful with a protocol, careful with a dose, careful with the follow-up question nobody else thought to ask. Then the resource you promised on air never shows up, and the same audience is holding both of those facts at once.

The gap is where those two things come apart.

Your promises are living inside the artifact

Go find every commitment you made in the last two weeks. You can't, and it isn't a memory problem.

Promises get made inside things. Inside a 58-minute recording. Inside a discovery call. Inside a thread that's now forty messages long. Inside a DM you answered from your phone in a parking lot. The commitment lives in the artifact, and the artifact gets filed the second it's over.

Your task list, meanwhile, was built from a completely different input. It holds the things you decided to do while sitting down and thinking about what to do. It's missing everything you decided while talking.

Run the audit. It takes ten minutes.

  • Your last episode or long recording. How many times did you say some version of "I'll link that"?

  • Your last three client calls. What did you say you'd look into?

  • Your last week of email and DMs. What did you say you'd send?

Count them. Then count how many actually went out.

The distance between those two numbers is the part you have never once looked at.

Catch it at the source, while the artifact is still warm

The fix is boring and it works. Sweep the artifact for the promise before you file it.

When a client uploads an episode, the pipeline reads the transcript and pulls every place the host said a version of "I'll put that in the show notes." Then it goes and finds the thing. The actual study, the actual link. The show notes get built with all of it already in place, and nobody has to remember anything.

You don't need my pipeline to start. You need the sweep. Here's the prompt I'd run on any transcript, call note, or long thread:

You are reviewing a transcript for commitments the speaker made.

Read the text below and extract every promise, explicit or implied,
that the speaker made to another person or to an audience. Include
anything they said they would send, link, share, look into, follow
up on, check, or include later.

For each one, return:
- The exact quote where the promise was made
- What is actually owed (be specific about the thing)
- Who is owed it
- What would have to happen for it to be delivered

Do not summarize the conversation. Do not include private intentions
the speaker voiced that created no expectation for anyone else.
Capture only what another person is now waiting on.

Transcript:
[paste here]

Run it on your last episode. Run it on the notes from your last three calls. The output is a list, and a list is a thing you can finish. Making the promise was never the hard part. Finding it again was.

Same principle, different artifact

Your inbox is full of promises too, and it's the one place they arrive faster than you can log them. That's the reason I built the Email Triage Agent. It reads what came in and tells you what's actually owed, so the follow-up stops depending on you remembering the message existed.

The kit is free. It's the setup, the prompts, and the logic I run on my own inbox.

Baldomero Garza — Find me on X, LinkedIn, Instagram, or book a 1:1.

Reply to everything. Edit nothing.

Your inbox is full. Slack is piling up. Client messages need a response yesterday. Typing thoughtful replies to all of it takes hours you don't have.

Wispr Flow turns your voice into clean, professional text you can send the moment you stop talking. Speak like you would to a colleague — tangents and all — and get polished output. Emails, Slack, LinkedIn, WhatsApp, whatever's open.

89% of messages sent with zero edits. Used by teams at OpenAI, Vercel, and Clay. Works on Mac, Windows, and iPhone.

If you found this issue helpful, please forward it to a friend or colleague who might benefit from it. Sharing is caring!

Did someone forward you this issue? Don't miss out on future insights — subscribe to the Ops & Om newsletter here!