In an AI world, judgment is the new superpower

Don't outsource yours (at least not yet)

👋 Hi, it's Greg and Taylor. Welcome to our newsletter on everything you wish your CEO told you about how to get ahead.

I've been pushing my team to use AI for everything for the last 18 months. I regularly respond to messages with, “Have we run this through AI?”

But lately, I've been frustrated when I get documents that feel mostly AI-generated with a few human edits sprinkled in. People even lead with phrases like “ChatGPT put this together…” or “AI says we should…”, which makes me batshit crazy.

These aren't lazy people. They're doing exactly what I asked them to do. But I don't want to know what AI thinks - I want to know what THEY think.

In trying to be more efficient with AI, we’re outsourcing what makes us valuable: our judgment.

And in a world where we can’t compete with AI on speed, we MUST compete on judgment - or we’re not really contributing anything at all.

Here’s my latest thinking on this - I’d love to hear yours too.

– Greg

Why judgment is your new superpower

AI is great at what I call “fast, tidy, and bland” – a standard analysis with obvious recommendations that gets 80% of the way there in 5 seconds.

Obviously there’s huge value to this. A “good enough” product is fine for a lot of knowledge work (e.g., summarizing a report) and speed matters if you’re trying to compete.

But right now, AI can’t do the 20% to make the output or recommendation REALLY specific to your business.

A strong recommendation requires context and insider knowledge – insights from what you’ve done before, risks that AI doesn’t know about, constraints your business is currently facing, strategic trade-offs that require knowing your customers and team, etc.

That 20% is where your value lies. It's what separates a great executive from a great prompter.

Here’s an example. Last week, Taylor was writing the job description for our ProfAI product manager. She spent 30 minutes up front writing down her POV (bullets) - the specific skills needed in our product manager at Section.

She gave that to AI, told it to make sure the JD also covered standard PM responsibilities, and went back and forth with AI over 15-20 chats to hone a tailored JD for Section.

She could have just asked AI for the first draft of that job description - and gotten a strong, generic description of a product manager. That’s the 80% “good enough” output, and it has value, because there are standard components to a PM job and it saves time.

But the 20% is making that JD relevant to Section - the ways we need this person to work, the specific complexities of this product, etc.

The goal should be using AI to get to that 80% faster so you can spend more time on the 20% that's uniquely yours. But too many people are stopping at the 80% and calling it done.

The tells that say “I’m outsourcing my judgment to AI”

There are two dead giveaways that someone has outsourced their thinking:

They say they did. Never say "AI says..." in a meeting or document. Ever. This is the fastest way for you not to be invited back to meetings. It's problematic on multiple levels: AI is often wrong or missing context, and more importantly, why do I need you if you're just going to be AI's spokesperson?

You wouldn't say "the data says we should do X" or "three customers say we should do Y” or “my colleague Jenna says we should do Z.” You take inputs, add context, consider resources and risks, and form a point of view. That's your job as a rising executive.

The formatting gives them away. There's a formatting “tell” that screams “AI wrote this” (and no, it’s not the em dash - I’m sick of seeing that on LinkedIn).

It’s tidy formatting – bullets, bolding, headers – and bland prose. There’s value in the formatting - it is one way to eliminate friction for the user. But when I see it, I know to read more closely - and I often find generic thinking underneath. Then I delete the report or doc.

How to actually collaborate with AI

Here's the process that works:

Start with your own thinking. Do the initial work – crunch numbers, review data, form a quick hypothesis. This can be fast – 30-60 min depending on the scenario. But take a beat and write it down. Don’t agonize over it but be prepared to show me your first take if asked.

Then work with AI. Use it to expand your thinking, expose blind spots, pressure-test assumptions. Ask it about unintended consequences you might have missed. Ask it about different approaches to get the same outcome. (Side note: Taylor is teaching a workshop on how to do this part next week – if you want to join for free send us a note).

Form your updated point of view. It might be the same as your initial take, or AI might have shifted your thinking in minor or major ways.

Critical step: If your POV changed significantly, go back and check AI's work. Verify assumptions, question the data, examine the reasoning. AI is a thought partner, not a source of truth. You would never blindly take the recommendation of a mentor or advisor (who doesn’t work in your business) and just follow their directions. So why do this with AI?

Own your final recommendation. Ask yourself: Do I believe in this? Do I have conviction? If my boss pushes back, will I fold and blame AI, or can I engage in productive dialogue?

My advice

Every CEO is telling their team to use AI more, faster, for everything.

So when I say “make sure it's your own thinking,” it might feel like I'm asking you to pretend AI didn't help, or to redo work that AI already did well.

But there's a difference between using AI as a tool and using AI as a crutch.

Using AI as a tool means using it to pressure test your thinking and make it stronger. Using AI as a crutch means outsourcing your thinking entirely, then presenting AI's conclusions as strategic insights.

If you've been over-relying on AI and find yourself in a meeting getting pushed on something you can't defend, fess up fast. Don't blame the machine. Just say you rushed the analysis, over-relied on AI and will do better next time. You get one “my dog ate the homework” card.

Within a year, using AI won't be a differentiator – everyone will be doing it. Your value will be in your judgment, your context, your ability to synthesize inputs and make decisions.

Don't let AI make you replaceable. Let it make you faster.

Have a great week,

Greg and Taylor