The best place to start with AI isn't the technology. It's a problem you know well enough to care about solving.

A few weeks ago, I wrote that in the age of AI, everyone is a builder.

I still believe that.

Actually, after spending an hour last week talking with people about what it means to build in this new era, I believe it even more.

But I realized there is a second question hiding underneath the first one.

If I'm a builder...

What do I build?

I think this is where AI can become overwhelming really quickly.

Open any AI platform, and there are seemingly endless possibilities. Agents. Models. Coding tools. Automations. Workflows. Every week, another tool promises to change how we work.

It's tempting to start there.

I don't think we should.

Don't Start With AI

One thing that stood out to me in our conversation recently was how quickly we wanted to get to the technology.

What tools should I use?

Do I need to know how to code?

Which model should I build on?

How do agents work?

Those are useful questions. Eventually.

But they aren't the first question.

The first question is much more human: What problem do you understand deeply enough to solve?

That changes the conversation.

Because suddenly the pastor who has never written a line of code but has spent ten years watching new people fall through the cracks at his church has something incredibly valuable to bring.

So does the nonprofit leader who knows exactly where her team loses hours every week.

So does the college student who sees a problem on campus that everyone else has simply accepted.

So does the developer who has experienced something frustrating a thousand times and finally thinks, “There has to be a better way.”

You don't need to go looking for an "AI problem."

Look for a human problem.

The Problem You Know Might Be Your Advantage

This is the part we underestimate. We tend to believe expertise means knowing how to build the solution. Sometimes expertise means knowing the problem better than anyone else.

You know where people get stuck.

You know the workaround everyone uses because the actual process doesn't work.

You know the spreadsheet someone manually updates every Thursday.

You know the person who gets forgotten.

You know the question people keep asking.

You know the thing everyone on your team complains about but nobody has had the time or resources to fix.

That knowledge matters.

In the previous article, I argued that what's missing from this moment isn't simply more technical skill. It's presence. We need the people who actually understand these communities to participate in what gets built.

This is what that looks like in practice.

Start with what you know.

Then Get Uncomfortably Specific

"Help churches" isn't a problem.

"Use AI for discipleship" isn't a problem.

"Improve human flourishing" definitely isn't a problem.

They're ambitions. A buildable problem is smaller.

Our church has 40 new people fill out a connection card every Sunday, and our staff can't personally follow up with all of them by Monday.

Now we have somewhere to start.

Or: Our food pantry doesn't know what it will run out of until a volunteer notices the shelf is empty.

Or: Students join our campus ministry during their first month of college, but we don't know who is quietly disappearing three weeks later.

Specificity gives you something AI hype never will: A way to know whether what you built actually helped someone.

Build With People, Not Just For Them

This may have been my biggest takeaway from the conversation. The strongest builder isn't necessarily the person who can do everything. It may be the person who knows what they can't do and finds someone who can.

A ministry leader brings context.

A developer brings technical depth.

A designer understands behavior.

A theologian sees questions the rest of the team might miss.

Someone who has actually experienced the problem brings something no amount of research can replicate.

Put those people together and the solution changes.

This is why I keep coming back to the word together when I think about building in this era. AI lowers the barrier for an individual to make something, but that doesn't mean the future should be built individually. Quite the opposite. The easier technology becomes to build, the more important it becomes to bring different kinds of wisdom into the building process.

Your First Version Should Probably Be Bad

I think some of us need permission to hear this. The first thing you build probably won't be very good.

Build it anyway.

We're used to seeing finished products. We see the app after it launches. The company after it raises money. The tool after thousands of people are using it. We don't see version one, and version one is usually awkward.

The workflow breaks.

The prompt doesn't work.

The button is in the wrong place.

You realize halfway through that you misunderstood part of the problem.

Good. Now you're building.

The tools available right now allow us to move from idea to prototype faster than almost any other moment in history. That doesn't mean we suddenly get everything right the first time. It means we can be wrong faster, and then learn. And then, build again.

The Goal Isn't to Build More AI

This matters to me. The goal of this moment cannot simply be to put AI into everything. More technology is not automatically better technology, and more automation is not automatically progress.

The question underneath all of this has to remain:

Does this actually help people flourish?

Does it give a pastor more time to sit across from someone who needs them?

Does it help a nonprofit serve someone who might otherwise be missed?

Does it strengthen community instead of replacing it?

Does it help someone find truth?

Does it give people more capacity for the deeply human things only people can do?

If the answer is no, maybe we shouldn't build it. The ability to build more makes discernment more important, not less.

Start With the Thing You Can't Stop Thinking About

So yes. You're a builder!

We've established that; now stop worrying about whether you know enough about AI to begin. Look around.

What frustrates you?

Where are people being underserved?

What takes far too long?

Where does someone consistently fall through the cracks?

What problem do you understand because you've actually lived it?

Start there. Name it clearly. Make it smaller. Find people who understand parts of it that you don't. Build something. Let it be bad. Learn. Build it again.

Because the opportunity in front of us isn't simply that AI allows more people to build technology. It's that the people closest to the problems finally have the ability to help build the solutions. And that might change what gets built altogether.