What's Possible With AI Comes Down to How Big You're Willing to Think

If you want to know what's possible with AI in your business, the honest answer is that the limit sits with the size of the thing you are willing to ask for, well before it sits with the technology. Most people run a small experiment, get a decent result, and quietly file AI under "helps me write emails faster." The tools were capable of much more the whole time.

I watched this happen in real time on a group call. I asked one of the members what they could create with AI that would make a personal goal of theirs inevitable. They answered fast and honestly: "I'm going to be honest. I'm not really in that space of creating. I'm more of a doer." Then they turned to another member on the call, told them to build it, and offered to pay a monthly fee for it once it existed.

So I asked a direct question. Are you aware you can create an app using Claude Code? They said yes.

That is the whole thing. They knew. The knowledge was already there and it changed nothing, because they did not count themselves as someone who builds. The ceiling sat in how they saw themselves, and the tooling had cleared out of the way long before.

What's possible with AI for someone who isn't technical?

More than the people asking usually believe, and the gap has widened fast.

Two days before that call, my four and a half year old son was in the middle of a sleep regression. He had been great at going to bed until the last two weeks. Then it turned into "I'm afraid of the dark," "my tummy hurts," "I miss mama."

So I built him a game. He picks what he wants to send mama, a heart, and then he squeezes a panda bear on screen. It tells him "Panda has you." Then "Mama always comes back." Then "You did it."

Then I gamified his whole nighttime routine. PJs, brushing teeth, potty, book, room check, comfort squeeze, lights out. Every step completes and he gets something for it. There is a duck who runs bubble breaths with him. Breathe in, breathe out, pop it.

A bedtime app is a small thing. I showed it on that call for one reason. The member who had just said they were not in that space of creating got to look at something finished and personal that I had built days earlier, and then I asked them straight out what their version could look like. That is a different question than whether it can be done at all.

That is the real answer to what's possible with AI for a non-technical person. If you can describe it clearly, you can very likely have it. The people getting the most out of this ask for bigger things than everyone else does.

If you want the ground-level version of how that works, I wrote about what Claude Code actually looks like for people who don't write code.

Why do most people underestimate what's possible with AI?

Because they calibrated their sense of what is buildable back when building was expensive, and nobody sent them a memo when that changed.

Here is how I say it to the people I coach.

The bigger you dream and think, the more AI can support your vision. We are only limited by the ceiling of our own capacity. Our ability to think and dream bigger is a skill, and it is trainable like any other.

The underestimating shows up in a specific way. People are fluent and generous when they imagine what is possible for somebody else, and they go quiet when the same thing is aimed at them. On that same call we ran an exercise where one person speaks a big goal for thirty seconds and everyone else spends two minutes on a timer saying "what if" at them. What if it happened faster. What if the money is already waiting. What if this is the easiest thing you do all year.

Everyone was excellent at giving it. Receiving it was harder. One member said it plainly afterward: "I enjoyed speaking it, no issues, but on the receiving end, in full transparency, I had very quick fleeting moments. Like really?" Then they caught their own objection mid-sentence. "If they can speak this in existence, why can't I?"

Another described "a little uncomfortableness of like, can that really be the truth?"

It is easier for us to give than to receive. That asymmetry is expensive, because your ambition is the input AI is working from. A modest ask produces a modest result, and then the modest result gets treated as evidence about the tool.

How do I start thinking bigger about what AI can build?

You borrow other people's belief until yours catches up, and you make the borrowing deliberate instead of accidental.

The thirty second exercise I described above works because of how it is structured. The person receiving says nothing. They sit there and take it. No defending, no explaining why the timeline is unrealistic, no managing anyone's expectations. Then everyone claps and they let it land.

Watch what that does. One member spoke about buying a second property they could rent out. I asked what if this happens in the next thirty days. Their entire answer was one word. "Done."

That is a ceiling moving in real time.

Another member described the software they want to build and told the group it would be at version two or three by the end of the year. I stopped them there, because they had skipped a level without noticing. How do you know when version one is live? Not done. Live. Version three has to have version one and two underneath it, so what does creating version one actually look like?

People protect themselves from starting by describing a mature version of the thing. Define what live looks like and the work becomes ordinary again.

This is the same muscle I wrote about in how to dream bigger and make it real, applied to a moment where the tools finally caught up to the dreaming. It also sits close to the mindset shifts that separate entrepreneurs who get results from AI.

How do I find out what's possible with AI in my own business?

Stop researching it in the abstract and put a real problem in front of it that you actually care about.

The reason the bedtime app was worth showing is that it was specific and personal and finished. It needed no briefing on AI capability. It was a thing that existed. Finding out what's possible with AI in your own business works the same way. You get one honest data point by building something small and real, and that data point recalibrates everything after it.

Three questions that get you there faster than reading another list of tools:

What is the thing I keep saying would be great if it existed? That sentence is usually pointing at a build you have already talked yourself out of.

What did I hand to somebody else because I decided I was not the person who makes things? That was the exact move on our call, and it is worth catching in yourself.

What would I attempt if the build were free? Answer that one honestly, because it is now much closer to true than your instincts are pricing in.

The people I see getting outsized results are running this loop constantly rather than annually. High performers use AI differently, and most of the difference is the size and frequency of what they ask for.

What should I build with AI first?

Build the version of something you already want that is small enough to finish this week.

Skip the platform and skip the company-wide system for now. The bedtime game was small enough to hand to a four year old and finished enough to put in front of a room full of business owners. Small and finished beats large and theoretical, because finished is what recalibrates you.

I told the member who had passed on building to consider their own version. They have a personal goal they have been carrying a long time. So build the app for that. Make it yours, make it fun, log the journey, put whatever would make you open it every single day. The app is the byproduct. The real output is becoming a person who builds the thing instead of describing it.

---

How to put this into practice

1. Write down the thing you keep saying should exist. One sentence. Do not evaluate it yet.
2. Define version one live, not version one done. What is the smallest version that a real person could actually use this week?
3. Build that version yourself. Even if it is rough. Especially if it is rough. The output matters less than the identity shift.
4. Say the bigger goal out loud to people who will build on it. Speak it to a group that will add to it rather than stress-test it. When you speak something to another person, it gives life to it.
5. Make it a daily rep. Pick one creative rep you do every day as a gift to yourself. A friend of mine writing their first book gave themselves one hour of writing a day. People write a book and nine months later, there it is. Building a daily AI practice is the same principle pointed at your business.
6. Look at the goal more than once a year. In The Power of Consistency, Weldon Long makes the point that most people do not achieve their dreams because they do not spend enough time in them. Most people look at their dreams once a year. The best in the world look at theirs once a day.

The real ceiling

There are 122 days left in this year. That is enough time to build the thing you have been describing.

The member who started the call saying they were not in that space of creating had every piece of information they needed before we ever got on the call. They knew the tool existed. They knew what it could do. What they were missing was a picture of themselves as the person who uses it that way.

When I asked at the end what was alive for them, one member said this: the power of possibility, and the only limitation, the only thing that would slow me down, is myself.

That is the answer to what's possible with AI. Everything you are willing to actually ask for. The tools stopped being the hard part a while ago. What is left is whether you let yourself want something big enough to be worth building.