Answers.Founder questions, by Humble

I have an idea but can't code. What's the next step?

Code stopped being the hard part. The next step is a person who has the problem, and you don't need a developer to find one.

2 min read · Updated 3 Nov 2026

The short answer

If you have an idea and can't code, your next step is finding someone who has the problem. Write the idea in one sentence, name the first customer you suspect, and talk to five of them about the last time it happened. When you do need software, AI tools now let non-technical founders build a first version, so code is rarely what's holding you back.

The old version of this question had a real answer: find a technical cofounder, raise money, or learn to code. In 2026 those are rarely the first move.

Code isn’t the bottleneck any more

In June 2026, Lovable told TechCrunch that people on its platform start a million new projects a week, and that its users are “primarily non-technical” (TechCrunch). A non-technical founder can now put a working first version on the internet without writing code.

That changes what the next step is. Building has become cheap, so the expensive mistake is building the wrong thing for the wrong person, at any skill level.

Your next step: one sentence, one person, five conversations

  1. Write the idea in one sentence. Who it’s for and what changes for them. “An app for small businesses” doesn’t count. “A Monday report that pulls four spreadsheets into one for ops leads at 20-person agencies” does.
  2. Name the first customer you suspect. A role and a situation, specific enough that you could go and find one this week.
  3. Separate what you know from what you’re guessing. Most ideas at this stage are almost all guesses, and that’s fine as long as you label them.
  4. Talk to five people who match that customer. Ask about the last time the problem happened, what it cost them and what they did about it. Leave the idea out until the end.
  5. Pick the one question that matters most. Usually it’s “will they pay for this?” Your first test exists to answer it.

When you do need software

Build the smallest version that lets one real customer get the result, and only once they’ve said they want it. For many ideas the first version can be done by hand, with forms, spreadsheets and email, before any app exists.

If you build it yourself with AI tools, get someone technical to look at it before you take money or store customer data. That review costs far less than fixing a problem after customers find it.

What to skip for now

  • A logo, a brand and a website. Useful later. None of them tells you whether the idea works.
  • Hiring a developer or an agency. Do it when you know exactly what the first version must do, and who’s waiting for it.
  • Looking for a technical cofounder. Customers you found yourself are the best way to attract one, so find them first.

The Humble Edge · Sunday email

Get one of these every Sunday.

One founder problem and one thing to do before Monday.

Questions

Should I learn to code first?

Not before you know someone will pay. Learning to code can take months; finding out whether the problem matters takes a week of conversations.

Should I hire a developer or an agency?

Only once you know what the first version must do and who it's for. Paying someone to build an untested idea is the most expensive way to test it.

Can I build it myself with AI tools?

Often, yes, for a first version. Lovable told TechCrunch in June 2026 that its users are primarily non-technical and start a million new projects a week. Have someone technical check it before you take payments or store customer data.

Do I need a technical cofounder?

Not to start. A cofounder helps most once there's something worth building, and customers you found yourself are what attract a good one.

Try something

While you're here

Next step

Is your idea worth 30 days? Find out free.

Five questions about your idea and a straight answer. It's week zero of Humble's 30 days, and it costs nothing.

More in idea