The Brief Is a Decoy

And here’s the uncomfortable thing about your AI sessions. The most valuable thing you got from one today probably wasn’t what you asked for.

It was the other thing. The tangent. The thing that started as a throwaway observation and quietly turned into something you should have captured but didn’t — because you stayed on task. Because you were disciplined. Because you’d been trained your whole career to finish the brief and move on.

That discipline is destroying your best AI outcomes.


Knowledge entrepreneurs are efficient. That’s not incidental — it’s how they survived. You built a business out of your thinking, which means every hour has always been on the meter. You learned to kill tangents. You learned to close loops. You learned to protect the deliverable. And for twenty years, that was the right skill.

The problem is, AI doesn’t work like a project. It works like a conversation. And conversations produce value in the margins, not just the middle.

The brief tells the AI what you already know you need. The tangent reveals what you didn’t know you were about to figure out.

When you cut the session the moment the deliverable is done, you’re leaving that on the table every single time.


I’ve watched this happen three times in my own work in the past year, and each time the pattern was identical. The brief was a decoy. The real artifact was hiding just past it.

First one: I needed a quick X post about fixing AI slop in content output. That was the brief. We wrote the post, done. But somewhere in the middle of that session, we started riffing on what “quality” actually means in AI-generated content — how do you score it, how do you track it over time? Fifteen minutes later, Claude and I had sketched out an eval loop with a running quality scoreboard. Nobody planned to build that. The post was fine. The quality-gate system is now running in production.

Second one: I asked for a rewrite of some content in the style of T. Harv Eker — just for fun, to see what happened. The brief was a one-off experiment. But the process of doing it forced us to name what made Eker’s voice work — the contrast triplets, the wealth files, the declaration structure, the seminar rhythm. What we ended up with wasn’t a rewrite. It was a full forensic voice profile. We called it Grounded Challenger. It lives in the skills library now, permanently reusable. Every time a piece of content needs to channel that register, I load it. That exists because I didn’t close the session when the rewrite was done.

Third one: I asked a simple classification question. Yes or no — is this an agent or a skill? I just needed an answer. But the question wouldn’t sit still. Claude pushed back, kind of. We started trying to name what the actual distinction was, and then what the distinctions were, and somewhere in there we landed on four types: command, orchestrator, skill, ambient agent. That four-type taxonomy for ambient intelligence is now the architecture I use to explain everything I build. It came from a yes/no question I asked on the way to something else.

Same pattern, three times. The brief started something. The tangent is where I actually went.


So why do we keep killing the tangent?

Part of it is just conditioning. Knowledge entrepreneurs have spent decades being rewarded for staying on task. You got hired for follow-through. You built your reputation on delivering. Every coaching certification, every consulting engagement, every course you created — it trained you to treat focus as a virtue and tangents as a failure of discipline.

And then the AI productivity culture came along and made it worse. “Better prompts, faster results, tighter sessions” — that’s the whole genre. Productivity content for AI users is obsessively about efficiency. Get what you need and close the tab. We turned efficiency into the whole point when efficiency is just the vehicle.

The other part is psychological. The brief is a contract. You wrote it down, or you said it out loud, or you just held it clearly in your mind — and now it defines what success looks like for this session. Finishing the brief feels like winning. Wandering off it feels like losing. Even when the wander is where the gold is.

The brief is doing something the project management world calls “scope creep prevention.” In an AI session, that’s not a feature. It’s a cage.


So what do you actually do? You don’t abandon the brief — you finish it. Then, before you close the session, you ask one question:

“What did we almost build that we didn’t?”

That’s it. That’s the whole protocol.

Give it fifteen minutes. Not an hour, not a new session — fifteen minutes, right there, before you close the tab. Most of the time nothing comes of it. Every few sessions, something does. When it does, you capture it as its own artifact — not a footnote in the original file, not a note buried in the chat — its own named document, its own starting point.

The Grounded Challenger voice profile lives in its own folder. The eval loop has its own repo. The four-type taxonomy has its own writeup. They exist as real things because we named them and gave them a home instead of letting them dissolve back into the session.

Here’s what the protocol looks like in practice:

Finish the brief. Don’t skip steps to get to the tangent. Deliver what you came for.

Before closing, run the one question. “What did we almost build?” Say it out loud to the AI if you need to. Let it reflect back what the conversation actually produced.

Give it fifteen minutes. Not “maybe next time.” Right now, while it’s live.

If it produces something, save it separately. Name it. Give it a home. A fragment you didn’t name disappears. A fragment you named is the seed of the next thing.

That’s the whole move. It doesn’t require a different mindset. It requires one question and fifteen minutes you’d usually spend closing tabs.


The brief got you started.

The tangent is where you’re going.

Next time you’re in a session and something unexpected surfaces — some idea that isn’t what you asked for, some framework you didn’t plan to build — don’t execute discipline. Execute curiosity. You can always come back to the brief. The tangent only lives for the next few minutes.

Let’s serve people, do good, have fun and make money — abundantly. Namaste.