The answer was one sentence: a project is a place, a skill is a procedure. When I ran that sort across my folders, almost everything I'd written down was a procedure, which means I'd written most of it more than once. One subject-line rule, fourteen places. Every improvement cost fourteen edits, so improvements quietly stopped happening.

👉 More folders than you can keep current? Let's pick the one system worth fixing first — book a 1:1:

KEY TAKEAWAYS

1. Sort every documented line in four seconds

2. Duplication looks like a well-organized archive — which is why it survives for years.

3. Write the job once, write the client once, let the folder become a pointer.

4. Duplicate on purpose

5. Map every folder before you migrate one

TIMESTAMPS

00:00 START

01:15 – The boring logistics question that started all of this

03:05 – Moving a personal account over to a team account

04:40 – a project is a place, a skill is a procedure

07:00 – Why my well-organized folders were an archive, not a system

09:20 – The four-second sort

11:30 – The subject-line example: one rule, fourteen places

15:10 – what plugins actually are

17:25 – The three-move fix

20:00 – Nine pages down to twelve lines

24:15 – Don't migrate folder by folder

26:30 – Your 15 minute assignment

RESOURCES

1. Book a 1:1: https://calendly.com/thebaldogarza/60min

2. The Email Triage Agent Kit (free): https://newsletter.thebaldogarza.com/products/the-email-triage-agent-free-kit

3. Issue #20 — "My Free Thing Is a Product. My Paid Thing Is a Sentence.": https://newsletter.thebaldogarza.com/p/free-thing-product-paid-thing-sentence

4. Episode 3 — "Delegating With AI Agents Was the Easy Part.": https://youtu.be/tzUOY84HTRE

🔗 Connect with Baldo Garza:

Newsletter: https://newsletter.thebaldogarza.com •

Instagram: https://www.instagram.com/thebaldogarza/ •

LinkedIn: https://www.linkedin.com/in/thebaldogarza/ •

X/Twitter: https://x.com/thebaldogarza •

Facebook: https://www.facebook.com/profile.php?id=100093480696106 •

YouTube: https://www.youtube.com/@thebaldogarza •

Website: https://www.thebaldogarza.com

She was running the work on day one, because I didn't hand her a task — I handed her an agent. The judgment was already in the system. The handoff wasn't "here's how I think about show notes," it was "here's the system, go run it." The system is the training.

Then I lost a week moving my whole AI setup from a personal account to an organization account. Project memory doesn't transfer. Context built over months doesn't copy and paste. Connectors get approved at the org level and then again per person. Every tool I'd ever set up quietly assumed exactly one human would sign in. Everything I know how to do was written down — almost nothing about how I work was set up to be shared. Those are two different kinds of readiness, and only one of them gets talked about.

👉 The system my assistant runs on uses the same rulebook as my inbox agent — guardrails baked in. Grab it free: https://newsletter.thebaldogarza.com/products/the-email-triage-agent-free-kit

KEY TAKEAWAYS

• Hand over the agent, not the task — when the judgment lives in a system, there's nothing to apprentice.
• The steps you skipped aren't the complicated ones, they're the invisible ones. Screen-record yourself once and make an SOP out of it.
• Your infrastructure is the bottleneck, not your knowledge. You only find that out when someone else tries to use your setup.
• Write your AI's context and memory out to a doc as you go — it doesn't do that for you.
• The signal a hire is working isn't hours. It's whether they hand you things you didn't ask for.

TIMESTAMPS

00:00 START – Welcome and episode overview

02:20 – Documentation isn't what's standing between you and help 05:00 – Why the handoff took almost nothing: an agent, not a task

09:20 – The one gap: the invisible steps nobody writes down

12:10 – Where the week went: personal account → organization account 14:40 – Project memory doesn't transfer, context doesn't copy-paste 17:30 – Connectors, logins, and the one-person assumption

20:00 – Two kinds of readiness, and only one gets talked about

22:30 – The real signal: she flagged something that wasn't her job

26:40 – A business that only runs because you're in it is a job you can't quit

🔗 Connect with Baldo Garza Newsletter: https://newsletter.thebaldogarza.com Instagram: https://www.instagram.com/thebaldogarza/ LinkedIn: https://www.linkedin.com/in/thebaldogarza/ X/Twitter: https://x.com/thebaldogarza Facebook: https://www.facebook.com/profile.php?id=100093480696106

So I built it a memory in three layers — the chat where work happens, the project instructions that hold how it behaves, and a running history doc in Google Drive that logs every campaign. Then memory bit me. This episode is the day the agent told me I was wrong, explained why, and then admitted it had been wrong first — because it applied a lesson from a different product's campaign to one it didn't fit. That's the failure mode nobody warns you about: cached conclusions become dogma. They sound like experience.

👉 The whole thing runs on the same free rulebook I use for my inbox agent — guardrails baked in. Grab it free: https://newsletter.thebaldogarza.com/products/the-email-triage-agent-free-kit

KEY TAKEAWAYS

• Without memory, your agent re-derives everything every chat — and you wrongly conclude the tool is broken.
• Build it in three layers: the chat (it dies — get lessons out first), project instructions (behavior), and a history doc (what you've learned).
• Memory has a failure mode: an agent will obey its own history instead of thinking. Cached conclusions become dogma.
• Make it re-derive: if it can't explain how a lesson changes future behavior, it's a note, not a lesson.
• You can't automate yourself out of the loop — your job is asking the obvious question everything in its memory has already stopped asking.

TIMESTAMPS

00:00 START – Welcome and episode overview

00:40 – The AI ad agent that kept forgetting

04:00 – The realization: I was the memory

06:30 – The three-layer memory system

13:00 – The ritual: make it re-derive the lesson

17:00 – The June campaign that plateaued at 2.2x

21:00 – Two errors: cached conclusion + the disconfirming number 26:00 – The result: 9.15x on the new ad, 3.12x blended

29:00 – Memory's failure mode: cached conclusions become dogma 34:00 – Get the free email-agent kit

🔗 Connect with Baldo Garza Newsletter: https://newsletter.thebaldogarza.com Instagram: https://www.instagram.com/thebaldogarza/ LinkedIn: https://www.linkedin.com/in/thebaldogarza/ X/Twitter: https://x.com/thebaldogarza Facebook: https://www.facebook.com/profile.php?id=100093480696106

What if your inbox had an employee — one that works three shifts a day, never calls in sick, and costs almost nothing?

In this episode I break down the AI email agent I built to triage my inbox at 8am, noon, and 4pm, and why it works because of trust and design, not technology. The big shift: the project isn't "get AI to do my email." It's writing down how I make email decisions so clearly that someone else could make them for me. AI is just that someone else.

👉 Get the exact system free with my newsletter — rulebook template, guardrails, scheduled-task prompt, and setup checklist (including multi-account forwarding): https://newsletter.thebaldogarza.com/products/the-email-triage-agent-free-kit

Prefer it done for you? I'll build it. Same link.

KEY TAKEAWAYS

• Discipline doesn't scale, systems do.
• The asset is the rulebook — a text file that says what to do with every email, trained one question at a time.
• Trust by design: the agent never deletes, never sends (drafts only), never touches settings, never guesses.
• Worst case is just an extra email in your inbox. Build a safe box, then let it earn scope.
• One hub beats five inboxes — forward everything into one main inbox; originals stay untouched.

TIMESTAMPS

00:00 START – Welcome and episode overview

00:40 – The inbox that runs your life

03:00 – Why inbox zero doesn't scale

04:30 – The real project: writing down how you decide

05:30 – The three pieces: rulebook, labels, scheduled task

06:30 – How a rule gets made

09:00 – The four guardrails 11:00 – Worst-case failure mode

13:00 – Real wins: receipts, a property closing, an 86-store lead

17:00 – The multi-inbox fix

19:00 – Get the free files

🔗 Connect with Baldo Garza

Newsletter: https://newsletter.thebaldogarza.com

Instagram: https://www.instagram.com/thebaldogarza/

LinkedIn: https://www.linkedin.com/in/thebaldogarza/

X/Twitter: https://x.com/thebaldogarza

Facebook: https://www.facebook.com/profile.php?id=100093480696106

In this short trailer, host Baldo Garza shares what this podcast is about, who it’s for, and what you can expect moving forward.

Ops & Om explores the systems, strategies, and conversations that help people build healthier businesses, healthier lives, and more freedom. Through solo episodes, expert interviews, and themed series, we’ll cover topics ranging from business growth, AI, marketing, and operations to health, recovery, community, and personal development.

If you’re looking for practical ideas, thoughtful conversations, and real-world ways to create more leverage and less chaos, you’re in the right place.

Subscribe now and join us for the first series: Health of Business.

Systems. Scale. Freedom.