What AI Can’t Decide for You

A lot of people are uneasy about where AI is headed.

They worry it’ll erase jobs.
Replace skills.
Turn years of experience into a button click.

I understand the concern. But from the very beginning, I’ve felt something else was more true.

Humans don’t always choose the fastest path, even when it’s available.

Sometimes we actively avoid it.

Today in 3 minutes or less:

✔️ Why the mess is necessary

✔️ Fast answers = Dangerous

✔️ The blueprint before the house

✔️ The allure of analog

 🫠 Why the mess is necessary

If speed were the only thing that mattered, the world would look very different.

Take marketplaces like Etsy built around handmade work.
Factories could produce the same items faster.
Cheaper.
More consistently.

Yet people still wait longer. And pay more.

Why?

Because some tradeoffs are worth it. And deep down, we know efficiency isn’t the same thing as meaning.

Look at relationships.

They’re slow.
Messy.
Unpredictable.

They require attention in boring seasons and difficult ones, not just when things are fun or convenient.

If efficiency were the goal, we’d design them very differently, but they don’t work that way.

They work because they’re built over time.

⏩️ Fast answers = Dangerous

I kept coming back to this idea during a workshop I ran this week.

With AI, it’s easy to confuse confidence with correctness.

Fast answers feel right.
Clear answers feel right.

That doesn’t make them right.

Someone still has to slow down and ask:
Is this accurate?
Is this good?
Is this something I’d put my name on?

That moment of judgment doesn’t disappear just because the output came faster. 

If anything, it becomes more important.

If I needed to translate a book from English to French, I’d absolutely use AI, I’d get an answer instantly. 

It would sound polished, confident.

But I’d still have a problem.

I wouldn’t know if it was good or garbage.

That decision, knowing what’s acceptable, what’s accurate, what’s worth trusting, that’s the skill.

Not typing faster, not generating more.

Judgment.

🏠️ The blueprint before the house

Here's a practical way to slow down the process: stop asking AI for answers. Start asking it to ask you questions.

Instead of jumping straight to the solution, change your first prompt to basically interview you first.

Old way: "Build me a house"
New way: "Ask me questions to create a blueprint we can use to build a house"

This forces the creation process to slow down because now AI has to thoroughly understand what you actually need before giving you output.

Garbage in, garbage out.

Asking it to interview me first helps ensure what's going in is much better.

I'm not asking it for a fast answer. I'm asking it to understand the problem…

Which gets us much closer to a great answer.

That small shift has had a bigger impact on output quality than any other prompting technique I've tried.

💿️ The allure of analog

This showed up for me in a quieter way recently.

My son discovered my vinyl collection.
Old records.
Delicate equipment.
A system that’s easy to mess up if you’re careless.

One wrong move can ruin the whole experience.

He loves it anyway.

The ritual.
The attention.
The process.

It’s slower.
Less forgiving.
Objectively inefficient.

And that’s exactly what makes it meaningful.

What this means for you…

Some things are valuable because they take time.

The future won’t just reward speed.
It will reward people who can pause.
Notice nuance.
And say, “This one is actually good.”

AI can generate.
It can accelerate.
It can assist.

But it can’t replace discernment.

That responsibility still belongs to us.

You don’t need to compete with AI on speed. That’s a race you were never meant to win.

Instead, shift your focus to the parts of the work that can’t be rushed.

Start here:

• Reframe AI prompts: Instead of asking for an immediate answer, start with "Ask me questions to better understand what I need." Let it interview you first.

• Slow down one decision on purpose: Pick something you'd normally move through quickly and take an extra pass. Ask: Is this actually good? Is it clear? Is it something I'd stand behind a year from now?

• Notice where confidence is doing the heavy lifting: Fast, polished answers can feel right even when they're wrong.

• Protect one "inefficient" habit: Reading slowly. Reviewing carefully. Listening without multitasking.

Because the people who will thrive next aren’t the ones who produce the most, they’re the ones who can pause, evaluate, and say, “This one’s worth keeping.”

To profiting more and working less. 💸 

-Matt

GOLDEN NUGGET OF THE WEEK 👑

95% of HighLevel users think the ChatGPT function is just the "basic" AI option.

They're missing something big.

While everyone else builds simple chatbots, the ChatGPT function can actually do something Conversation AI can't...

And it changes everything about how you should be thinking about AI in your agency.

Watch to see the difference (and why the "simpler" tool might be more powerful).

HLPT CREATOR NETWORK 👥

🔒️ LIMITED TIME: $1 GETS YOU IN

If you like this newsletter, you'll love what's inside HL Pro Tools. Margin Makers readers get access to something special: 4 weeks of full access for just $1.

🌟 24/7 white-label support for up to 100 sub-accounts: your branding, our team handling the tickets

🌟 10+ live coaching calls every week: marketing, vibe coding, funnels, strategy, and scaling with people who've done it

🌟 1,000+ article support library: updated daily, fully branded to you

🌟 Done-for-you snapshots, funnels, and trainings so you stop building from scratch

🌟 A community of agency owners who are figuring this out alongside you

HAVE A TOPIC YOU’D LIKE ME TO COVER?

Sweet! I’m all ears.

Here’s your VIP pass to shape the future editions of Margin Makers 👇

Submit your ideas here.

DID A FRIEND FORWARD YOU THIS NEWSLETTER?

First, give them a BIG thank you!

Then, subscribe here to get your free weekly dose of entrepreneur life hacks.