8

min read

How to Use AI Without Outsourcing Your Judgment

A smart-sounding AI plan hits a real team meeting—and collapses. Why “AI Strategy Theater” happens, and how to stay in charge of the plan.

Ben Gledhill

TL;DR

A smart-sounding AI plan hits a real team meeting—and collapses.
That’s AI Strategy Theater: it reads well and dies on contact.
The fix: bring what’s real first, then use AI to clarify and stress-test—not decide.



The new trap: a plan that sounds smart but isn’t yours

A small business owner walked into a team meeting with a 6-page “marketing plan” ChatGPT wrote in 30 seconds—printed, stapled, and slid across the table like it was settled truth.

It sounded brilliant—like something a high-paid consultant would deliver after two weeks and a big invoice. It had phases, metrics, fancy language, the whole package.

Then his team asked:

“Who is this actually for?”

And someone else said what everyone was thinking:

“What does any of this even mean?”

That’s when he realized he couldn’t explain the plan he was asking people to run.

AI Strategy Theater is when a plan reads well—and dies on contact with execution.

It costs you credibility twice: once with your team, and again with yourself.

Because a pretty plan is still a guess.

Why this happens (and why it feels so good at first)

The danger isn’t that AI lies.

The danger is that it sounds certain—even when it’s guessing.

Here’s how it happens:

  • You give ChatGPT too little to work with—no baseline, no constraints, no real buyer, no plain-English offer.

  • You treat the smart-sounding answer like it must be right. Smooth sentences make your brain assume solid thinking.

  • You forget that thin inputs produce confident but fragile outputs.

That’s how you end up running a plan you don’t actually understand.

Relief isn’t clarity.

The three ways people accidentally outsource themselves

You can spot this—not in the output—but in your behavior.

1) Borrowed Voice

You start talking like the internet.

You run a relationship-driven business and suddenly you’re saying things like “increase top-of-funnel volume” or “build a nurture sequence.”

Your team asks:

“What do you mean by that?”

And you realize… you’re not totally sure.

If you can’t say it out loud and understand every word, it’s not your voice.

Rule: If a phrase feels weird in your mouth, translate it until it sounds like you.

2) Borrowed Judgment

You can present the plan.

But you can’t defend the choices.

Someone asks:

“Why are we doing this first?”

You nod like you expected it.

Then your eyes drop to the printout. Your stomach tightens—not because it’s wrong, but because you didn’t choose it.

Someone else asks:

“What are we doing this week?”

And you realize the plan isn’t inside you yet.

That’s the signal: the plan isn’t inside you yet.

3) Borrowed Agency

This is when you stop deciding and let AI be the boss.

You ask it for your next move, then the next. You don’t lead—you follow whatever pops out of the chat box.

You’re not choosing. You’re defaulting.

Defaulting is how smart people slowly hand over the wheel.

That’s not strategy. That’s dependency.

The five questions your team (or future self) will ask

Before you act on any AI-generated plan—pause.

Can you answer these five questions in plain English?

  • Who exactly are we targeting?

  • What are we actually offering?

  • Why are we doing this first (and not the other ideas)?

  • Who owns this next week?

  • What metric tells us it’s working—and when do we change course?

If you can’t answer them by yourself, you don’t own the plan.

You’re running AI Strategy Theater.

The Caldr method: You stay in charge

Here’s the Caldr method in one line:

You bring what’s real. AI helps clarify and stress-test. You decide and act.

This isn’t adversarial. It’s a partnership with roles.

AI helps you think better. It does not get to drive.

Do This Tonight (15 minutes)

Want a plan that fits your business—and one you can explain to your team tomorrow?

Step 1 — Write 7 lines in plain English

  • Goal + time horizon: what you want (this week / 30 days / next quarter)

  • Baseline: what’s true right now (how work comes in—referrals, inbound calls, etc.)

  • Best-fit buyer: who hires you fastest and stays happy

  • Offer: what you sell in plain English

  • Constraint: what can’t break (time, quality, reputation, etc.)

  • One move: one thing you’re willing to do next

  • One weekly number: a metric you can track weekly

Step 2 — Run a Red Team Check

Ask ChatGPT:

What am I missing?
What would you cut?
What’s the smallest next step?

Step 3 — Package execution

Then say:

“Turn this into an execution plan for the next [TIME WINDOW]: weekly actions, an owner, one weekly metric, and a simple ‘change course if X doesn’t move by week 2’ rule.”

Example:

“If booked conversations haven’t increased by week 2, we drop everything except reactivation and run 10 past-referrer outreaches that week.”

(For many people, [TIME WINDOW] is four weeks. Use what fits.)

Step 4 — Answer the five questions

Before you do anything, answer the five questions above.

If you can’t explain your answers yet, tighten the draft until you can.

A quick over-the-shoulder example

You:

“I’m trying to grow revenue by 25% next quarter, and I want help building a plan that’s realistic and executable.
Let’s do this in two steps.
First, I’m going to give you the truth about my business so you’re not guessing.
Then we’ll work together to turn it into a plan I can actually run.”

You:

“Here are the seven things I’m trying to get clear on.
If anything is vague or missing, help me sharpen it before we plan.”

  • Goal + time horizon: Grow revenue by 25% next quarter.

  • Baseline: Most work comes from referrals and repeat clients, but weekly volume is fuzzy.

  • Best-fit buyer: GCs and developers under schedule pressure.

  • Offer: We help teams understand ground conditions early so they avoid expensive surprises.

  • Constraint: Capacity and delivery matter, but the real limit isn’t fully defined.

  • One move: Re-engage past referrers.

  • One weekly number: Conversations booked (unsure what “good” is yet).

ChatGPT asks clarifying questions about the baseline and capacity. You answer.

You run a Red Team Check:

What am I missing?
What would you cut?
What’s the smallest next step we could actually run next week?

ChatGPT challenges assumptions and simplifies.

Then you say:

“Turn this into an execution plan for the next 4 weeks: weekly actions, one owner, one weekly metric, and a simple ‘change course if X doesn’t move by week 2’ rule.”

You also ask for a 60-second explanation for your team and three tough questions they’ll raise.

A quick definition

A channel is just the path work comes in—referrals, repeat clients, partners, inbound calls, content, ads. The way someone ends up choosing you.

If the word feels weird in your mouth, translate it.

Don’t borrow vocabulary. Borrow clarity.

If you can’t explain a word to a smart adult in plain English, it’s not helping you lead.

Closing

If you take one thing from this, let it be this:

If you can’t explain it, it’s not your plan.

ChatGPT can sharpen your thinking. It can simplify your ideas. It can help you say what you mean.

But it cannot own your decisions.

That’s your job.

You don’t need another smart-sounding strategy.

You need one that fits your constraints, matches your voice, and you can run next week.

Copy/paste: the 7-line Human-First Brief (fill this in)

  • Goal + time horizon:

  • Baseline:

  • Best-fit buyer:

  • Offer (plain English):

  • Constraint (what can’t break):

  • One move (next):

  • One weekly number:

Mini-FAQ

What if I don’t know my numbers?

Use ballpark ranges. “We get 6–10 referrals a month” is plenty.

You’re not trying to be precise. You’re trying to stop guessing.

What if my business isn’t ‘marketing heavy’?

Good. This isn’t about tactics. It’s about choosing one real path work comes in—and clearing the fog.

What if I already have a plan?

Run it through the five questions. If it passes, great. If not, tighten until it does.

What if AI gives conflicting advice?

It will. Use it to surface options. Then decide based on what’s true, what you can’t break, and what you can actually run next week.


Powered by the Clarity Engine™

© All Rights Reserved, 2025 Caldr


We use a few cookies to keep Caldr running smoothly. Learn more in our Privacy Policy.

Powered by the Clarity Engine™

© All Rights Reserved, 2025 Caldr


We use a few cookies to keep Caldr running smoothly. Learn more in our Privacy Policy.

Powered by the Clarity Engine™

© All Rights Reserved, 2025 Caldr


We use a few cookies to keep Caldr running smoothly. Learn more in our Privacy Policy.