Three AI Fluency Tactics by Guest Writer Z Zorrilla

I wanted to write something tactical that people could actually implement, so here we go.

Quick background. I started my career in Seattle real estate for about six years. Did pretty well. Was on a top team for three of those and won top broker awards for two. Then I joined Zillow and met Scott there. Annoyed the shit out of him, but I think I held my own. They shipped me across the country to help build out a team.

From there, I moved from sales into product. Worked as a product manager and eventually a senior PM across a couple of tech companies. Then I got laid off, which led to starting a brick-and-mortar SLP business with my wife. I worked my way out of the day-to-day, got bored, and now I'm at Zapier.

Why should you listen to me? Honestly, you probably shouldn't. But AI gets talked about constantly where I work, and my job title has AI in it, whatever the hell that actually means, so I've picked up some things.

Alright, onto the tactics.

1. Understand the landscape

You can't really talk about AI if you don't have a basic sense of how it works.

LLMs, or large language models, are tools like ChatGPT, Claude, Gemini, Microsoft Copilot, Composer, Grok, and others. Most people interact with them through chat interfaces.

Agents are essentially automations that you let these models build for you. Think of them as assistants who follow instructions and execute tasks.

MCPs and SDKs are just ways to connect AI to the tools you already use. Things like Jira, Slack, or Salesforce.

A harness is what ties everything together, so you can actually use it in a way that feels similar to the SaaS tools you're already familiar with.

You don't need to go insanely deep here, but you do need to know what's what.

2. Build something

It doesn't matter what it is. Just build something.

An automation, an agent, a small app. The goal is to get your hands dirty and understand how things connect.

Simple trick. Pay attention during your day and notice what annoys you. That repetitive task you keep doing over and over. Then open your AI tool of choice and ask it how it would solve that problem.

Then go try to build it.

You'll learn more doing that once than you will by reading ten threads about AI.

3. Measure the impact

Last one, and probably the most important. Measure what it actually did.

If you build something that reads and triages your emails, track how many messages it takes off your plate. Some of you are sitting on 5,000 unread anyway, so there's room for improvement.

If it summarizes your calendar and prep for the week, think about how much time that saves and what you might have missed without it.

If you build a workout tracker and cancel MyFitnessPal to save five bucks a month, but then burn $200 in tokens building it, write that down too.

The point is, don't just build for the sake of building. Understand the tradeoffs and the value.

Closing

The AI wave feels a little scary, unless you're on LinkedIn, where everyone seems oddly calm about it. But it's coming either way.

You can ignore it, or you can get familiar with it.

Figure out how you fit into it, or figure out what it can't do and double down there.

Either way, you don't get to opt out.

So you might as well lean in.

by Z Zorrilla  ·  Connect with Z on LinkedIn →