​

There's a question most professionals aren't asking themselves yet.

Not "am I using AI?" but "what actually changed in my output because of it?"

Those are very different questions. The first one is easy to answer yes to. The second one is uncomfortable.

Spent this past week looking hard at my own workflows. The honest result: months of "staying current," watching demos, testing tools, and only one thing genuinely changed my output. Research sessions that used to take most of a morning now take about 45 minutes. That's it. Everything else was just open tabs and confident-sounding fluency.

That's the gap worth talking about.

Fluent vs. Leveraged

Fluency is being able to talk about AI tools in meetings, name the right models, and sound informed. Leverage is your output per hour actually going up.

Most people are chasing fluency. It's visible, it's rewarded socially, and it feels like progress. The problem is that fluency alone doesn't show up when someone asks: "what used to take you 10 hours that now takes 2?"

Three different people mentioned to me in passing this month that their companies have started tracking AI usage as a productivity signal. Not surveillance, a performance metric. The question showing up in reviews isn't whether you use it. It's whether anything changed.

The Framework That Actually Helps: Theory of Constraints

There's a principle from manufacturing called the Theory of Constraints, made famous by Eliyahu Goldratt. Simple idea: any system only moves as fast as its slowest point. Before you add anything new, find where things actually get stuck.

Dan Martell applied this directly to AI in a recent interview. His point: most people pick the tool first, then go looking for a problem to match it to. That's backwards. Find the constraint. Then find the tool that removes it.

Here's what that looks like in practice, the 20-minute version:

1. Open your calendar from the last two weeks. List every task that took more than 45 minutes.
2. For each one, ask a single question: does this require my judgment, my relationships, or my creativity? Or is it mostly processing, searching, formatting, summarizing?
3. Circle the second column. That's where your bottleneck lives.
4. Find one tool that addresses that specific friction. One.
5. Go deep on it for two weeks before touching anything else.

For most knowledge workers, the answer shows up in research, synthesis, first-draft generation, and meeting prep. Those are high-volume, low-judgment tasks that compound fast when removed.

What Tim Ferriss Said That Stuck

In a recent Q&A, Tim Ferriss was asked what human abilities are becoming more valuable in a world full of AI. His answer: offline informational advantage.

The edge is what you know that isn't online. Your judgment. The relationships you can text for narrow expertise. The room you've been in that nobody else has. None of that is in any training set.

The implication is real: learning tools is the floor, not the ceiling. The ceiling is what you bring to the prompt that nobody else can.

What to Do With This

Run the 20-minute calendar audit this week. Not as a project. Just once. See what shows up in the processing column and ask if any of it actually needs to be you.

That single question is worth more than another hour of watching demo videos.

If this is useful, it comes out every week at wyattbrocato.site. Nothing else in your inbox, just one framework worth keeping.

​