Think Slow, Act Fast: The Only Way to Get AI Right

Everyone’s in a rush to implement AI wherever and however they can. And it's awesome. But when you're talking about strategic use of AI—AI that delivers ROI and makes a positive impact on revenue—you don't want to rush the entire process. There is a time to move fast and a time to move slow.

I've pared down my theory on this to four words: think slow, act fast.

Go slow in the beginning, when you’re planning and strategizing with AI. Ask the right questions, make a plan for mistakes and failure, and look at the long-term impacts of your plan. Then, when it's time to execute (the fun part!), move fast. Get scrappy. FAIL HARD.

Here's a closer look at how I approach this.

Think Slow

When you've started exploring use cases for AI, begin by taking the time to think through the use case and strategize (if it's a quick win, it might be worth it to just dive in, but not for bigger projects). This is when you should take your time because it determines whether AI makes your team smarter or just busier.

Daniel Kahneman calls this System 2 thinking in his book, Thinking Fast and Slow:. This is the deliberate, logical kind of thinking. It’s the mode that forces you to pause and ask hard questions before you move.

Start there by asking questions like:

• What are we really trying to fix?

• How does this fit into our broader marketing strategy?

• Where could AI actually help—and where might it make things worse?

• What does “quality” look like once AI’s involved?

• Which tasks are best automated and which tasks are best left to our marketing team?

This is not the moment to rush. This is the moment to be painfully clear.

Too many teams skip straight to “implementation mode” because they’re anxious to show progress. They want to prove they’re doing something with AI. But that shortcut usually backfires. You end up with tools that sound impressive but don’t improve output—or worse, they quietly lower quality.

Slow thinking means mapping out your process from top to bottom. Figure out where AI can add efficiency without flattening creativity. Think about how it will change the way people work, how it might shift tone or voice, and how it could affect customers down the line.

You don’t need every answer, but you do need to sit in the questions for a bit. They will reveal gaps that would be difficult to correct later down the line.

Act Fast

Once you’ve done that work—once you know why you’re using AI and how you’ll ensure quality and results—that’s when you pick up the pace.

This is Kahneman's System 1 phase: quick, intuitive, and experimental.

Start building.

Test as you go.

Break things.

Fix them.

Build again.

The faster you reach your failure points, the faster you’ll figure out what actually works.

Don’t overthink this part. Just move. Try a tool for one workflow and then test it in another, prototype an automation in a spreadsheet, and see what happens. The goal isn’t perfection—it’s momentum with feedback.

The beauty of this stage is that speed helps you learn. You’ll find better prompts, smoother handoffs, and sharper results through iteration. Over-analyzing and tuning every detail before you move on won't help you here.

Act Fast Part B

Now, there’s actually a Part A and a Part B to “Act Fast.”

Part A is about building the framework. Get the flow and the steps in place that are required to get the broader output you want, formatted and delivered the way you want. Scope out and keep it simple. You’re just making sure the system works end to end and does what you expect, overall

Part B comes next. Once the framework is solid, that’s when you move into precision mode. You start to really dial in the prompts and inputs, get very detailed, and test test test until the output is high-quality and consistent. This part is still fast, but it’s focused speed—rapid experimentation to fine-tune every variable until you can rely on your results.

Part A gets you motion.

Part B gets you mastery.

Balance Urgent vs. Important

As all of us who have been in this industry more than five seconds understand, urgent problems always shout the loudest. The inbox, the campaign deadlines, the things that make us feel productive. But the important work—the systems that compound over time—rarely yell. We have to make space for it.

AI is no different. The quick wins feel good, but the real transformation comes from building for the long term. Set aside dedicated time each week for those bigger-impact ideas. If you don’t, you’ll stay stuck in “urgent mode” forever.

AI Summary

Here's your quick, AI-like overview of what we just talked about because everyone's just super into that these days:

If you rush your strategic thinking, your AI strategy will turn into a pile of half-built, stagnant workflows. If you over-analyze the doing, your good ideas will never leave the page.

Think slow so your direction is right. Act fast so you can learn and adjust quickly and find the best methods to achieve your goals.