The Decision You Think Only You Can Make Is Probably a Script

The video that gave away the trick

We screen candidates on video before we ever get on a call. Same questions, sent to everyone, answered alone, recorded, sent back. The idea is simple: hear the person before you meet them.

The videos started coming back identical. Same structure, same phrasing, same rhythm, applicant after applicant. Then one written answer arrived with the AI's own sign-off still pasted at the bottom, offering to draft more follow-up responses. Nobody removed it. Nobody even reread it.

That's not a story about cheating. It's a story about a script running so smoothly nobody was watching it anymore. Send question, get answer, check box, move to next candidate. The whole stage had gone on autopilot, and it took a pasted line of AI boilerplate to interrupt it and force anyone to actually look.

Now ask yourself where that's happening in your own calendar.

The mechanism: scripts run until something interrupts them

The underlying principle isn't mine, it comes from Chase Hughes' work on behavioral influence, and it's simple: people run scripts. Automatic loops of behavior and interpretation that fire without conscious attention. A script left uninterrupted keeps the brain on cruise control. Interrupt it, and the brain snaps into resolve mode, full attention, briefly open, hunting to close the pattern break. That brief opening is the compliance window: the moment a new frame can actually get installed instead of bouncing off the old one.

There are levels to this. A social script is the handshake, the expected reply, the interview answer everyone gives. A life script is deeper: "I'm not good with money," "marketing isn't for doctors," the story you've told yourself so long it feels like fact instead of habit.

Decision Script: the life-script layer applied to authority, a decision your role now makes automatically, out of habit dressed as necessity, not because today's version of it needs your judgment.

You have one. It's quieter than "I'm not good with money" and more respectable: "this crosses my desk because it always has." Every approval, sign-off, and escalation that lands on you by default, not by necessity, is that script executing. Nobody questions it, including you, because it never gets interrupted. It just runs.

The decision that "requires" you and the decision that has simply always come to you look identical on your calendar. Only one of them actually needs your judgment.

Where AI actually helps, and where it just performs the same trick on you

Ask AI to "build me a decision rights matrix" and you'll get back something clean, structured, and generic, the same output anyone in your position would get, because you fed it a request with no context. That's the order-window move: put a question in, wait for the item that matches it to drop out, walk away. It's the exact failure mode from the candidate videos. Different person, same prompt, same AI, same answer. You're not mapping your decisions. You're comparing one AI's guess about your role to another AI's guess about someone else's.

The move that actually works is the one I didn't expect to work. I needed Rob's section cut out of a Zoom recording and opened a session with Athena, my AI partner, just to ask how it's done. I wasn't asking for the work, I was asking for the instructions. Instead she went and found where Rob started talking, where he stopped, cut the video herself, and handed me the finished file. I didn't know she'd be able to do that. I'd asked a procedure question and gotten an outcome back.

That's the whole shift, and it isn't a one-time trick for cutting video or building a matrix. It's a standing change in how you talk to any AI from here forward: stop asking for the steps, start stating the outcome, and keep testing things you assume it can't do. Most people never make that switch because the procedure-asking habit is its own script, comfortable, low-risk, never wrong because it never really tries.

The move, and what it costs you at the same time

Do this with your decisions. Hand AI your real material, the last month of approvals, the meetings you were cc'd into "just in case," the sign-offs that landed on you without a fight. Tell it what you actually want to know: which of these needed my judgment, which just needed my signature. Let it argue back.

Here's the part that doesn't come after, it comes with it. If this is done honestly, it will very likely tell you that a meaningful chunk of what crosses your desk didn't need you. That's not a workload finding, that's a threat to a story you've been telling about your own value, that you're the bottleneck because you're needed, not because it's habit. The compliance window that opens the moment you feed it real context and let it push back is the same window that exposes this. You don't get the useful interruption without also getting told your presence was optional in rooms you assumed required you. That's not a side effect, that's the mechanism working.

There's scar tissue too. If you've ever been burned by releasing a decision too early, your current habit of checking everything is insurance, not necessity. The map won't know that history unless you tell it. Feed it the log without the context of why you tightened your grip, and it will recommend releasing exactly the thing you have good reason to still hold. So the instruction isn't just "hand over the log." It's hand over the log and the reason you built the guardrail, or the map will be honest and wrong at the same time.

Rank this correctly

Script, interrupt, resolve mode, compliance window, that's not background theory sitting next to the action. The compliance window is the exact moment you hand AI your real decision history with real context and let it argue. That's where the old frame, "I have to touch this," gets briefly open enough for a new one to land. Skip the real context and you never open the window, you just get the order-window answer dressed up as a matrix.

Know when not to do this. Irreversible calls, who you hire at the top, who you fire, the decisions that define what the company stands for, the map can show you the pattern but it doesn't get to make the call. Use it to surface the decisions running on autopilot that don't need you, and keep your hand on the ones where being wrong isn't recoverable.

The candidates gave away their script with a pasted line of AI text. I gave away mine by asking Athena for instructions and getting the finished file instead. Your decision log will give away yours the moment you stop asking for a template and start handing over the real material, the real reasons, and the real cost of being told you weren't as necessary as you thought. The only question is whether you're willing to read what comes back.