You have an idea and nothing else. Here's what to do first.

I started GrowthMentor in 2018 for a slightly embarrassing reason. A few years earlier I was head of growth at a hosting company in Athens, professionally lonely, with a head full of half-formed ideas and nobody to talk them through with. So I did something a little odd: I went on Upwork and hired freelancers not to do the work, but to get on a call and let me think out loud, so someone could tell me whether I was about to do something stupid.

That feeling, an idea and no one to pressure-test it with, is the real starting condition for most first-time founders. You rarely lack information at this stage. What you lack is a sounding board and the nerve to begin.

If that is where you are, idea or pre-MVP, here is what I would spend my time on, and what I would happily ignore.

Build something you genuinely care about

Start with the unglamorous truth: this takes years. The exciting part, the idea itself, lasts about a week. After that comes a long stretch of unanswered emails, features nobody uses, and slow Tuesdays where nothing happens and you keep going anyway.

Passion is what gets you through that stretch. I do not mean the motivational-poster kind. I mean finding the problem interesting enough that you keep poking at it after it stops being fun. GrowthMentor came out of my own itch. I wanted the thing for myself, which is why I was still at it three years in, long after the smarter financial move would have been to quit.

I am doing it again right now, for exactly the same reason. I ran four marathons in a single year and still could not break four hours, and no training app was going to fix that for me. What I needed was a conversation with someone who had run through the wall I kept hitting. So I am building RunningMentor, the same idea pointed at a problem I live every weekend. It is very early and I am still testing it, which means I am going through every step in this post myself as I write it.

The honest test is a boring one. Would you still want to work on this if it took three years and never once made you look impressive at a dinner party? If the answer is yes, good. If you are mostly excited about being a founder, pick a different idea.

Direction is the scarce part now

Here is what changed. A few years ago, building software was the slow, expensive part of starting a company. You needed money, or a technical co-founder, or both, just to get to something you could show people. That bottleneck is mostly gone. With AI you can get to a working prototype over a weekend.

When building gets cheap, the bottleneck moves to direction: knowing which problem is worth solving, and whether yours is one of them. The real risk at the idea stage now is spending six months building the wrong thing beautifully, and only finding out once it is done.

Get your idea down to one sentence

Before you build anything, get your idea down to a single sentence a stranger can understand: the problem you solve and who has it. Skip the clever tagline. Aim for the kind of sentence where the other person nods instead of squinting.

I just spent a week reading interviews with founders who got into Y Combinator, and one trait came up over and over: conciseness. What gets you in is being legible. The partners need to describe your business in one sentence to each other and have everyone instantly understand the bet.

If you cannot get it down to one sentence, that is a sign the idea is still fuzzy in your own head, and building will paper over the fuzz instead of clearing it. So do the obvious thing almost nobody does: say the sentence to ten people who actually have the problem. Watch their faces. The ones who lean in and start telling you their own version of the problem are your signal. Polite enthusiasm is not.

Suggested mentors

Mentors who are sharp on positioning and getting a fuzzy idea down to one clear sentence.

Vassilena (Vassy) Valchanova

Message-mining and clarity. Turning a fuzzy idea into one sharp sentence.

Book a call

Eden

Conversion copywriting and voice-of-customer. The exact words your buyers use.

Book a call

Vanhishikha Bhargava

Positioning and go-to-market for early-stage products.

Book a call

Prototype fast, and why I reach for Claude Code

Once a few people lean in, build the smallest real thing you can put in front of them. The point is to learn something, so keep it cheap. Sometimes the right prototype is a landing page that explains the product with a button that measures whether anyone clicks. Sometimes it is you doing the work by hand behind the scenes while the customer believes it is software. Sometimes it is a rough working app.

For that last one, the tools have gotten genuinely good. Lovable, Replit, Bolt, and v0 can turn a prompt into a working app in minutes, and for a quick test that is often all you need. The catch is that your code lives inside their platform. Fine for something disposable, and it gets awkward the moment you want to take it elsewhere.

My own default is Claude Code. You get the same speed, and you keep a real codebase on your own machine that you own outright. Nothing is trapped in someone else's system. You can keep building on it for years, hand it to a developer when you finally hire one, or move it wherever you like. I am not a technical person, and I ship production code this way, which should tell you how far the floor has dropped.

Whatever you reach for, keep the prototype's job in mind. It is there to teach you whether you are onto something. It is not the finished product, and treating it like one this early is how months disappear. Build it, get it in front of your ten people, and pay attention to what they do, not what they say.

Suggested mentors

Mentors who actually build with AI, and can help you get a prototype in front of people fast.

Michael Taylor

Prompt engineering and building with AI. O'Reilly author on the subject.

Book a call

Sean Weisbrot

Shipping MVPs fast with AI tools like Cursor and Claude.

Book a call

Kosta Panagoulias

Bootstrapped two SaaS products from zero. Building lean and shipping.

Book a call

Talk it through with someone who has done it

Everything above is easier with someone in your corner who has already done it. This is the part I underrate the most when I talk to first-time founders, probably because it is the exact gap I started GrowthMentor to close.

A good mentor at this stage is someone you can bring the whole mess to in a single call. Is this problem real, or am I in love with my own solution? Which of these three tests should I run first? Should I build this in Claude Code, or just fake it by hand for a month? Am I kidding myself, and should I keep going? Those are not questions you want to be answering alone at 1am, and they are exactly the ones someone two steps ahead of you can settle in twenty minutes.

It is a different thing from advice. Advice is everywhere, and most of it is confidently wrong, because the person giving it is generalizing from their one story. What helps is a specific person who built a specific thing, looking at your specific situation, and telling you which of your ten worries is the one that matters.

It is not a one-off, either. Membership is unlimited calls with any mentor in the network, every one included.

Your bookings

$0 per call

Vassilena Valchanova

Positioning · Tue 10:00

Booked

Michael Taylor

AI & prototyping · Tue 1:00

Booked

Kosta Panagoulias

Zero to one · Wed 9:30

Booked

Tina Louise

Go-to-market · Thu 11:00

Booked

Daniel Johnson

Finding PMF · Fri 2:00

Booked

6 calls this monthadded to your bill: $0

Browse all mentors

A lot of it is just confidence

Here is the part nobody likes to admit.

At the idea stage, you are rarely blocked by a lack of knowledge. Almost everything you need to know is a search or a prompt away. What blocks you is nerve, the unspoken fear that you will commit, tell people, put it out into the world, and watch it not work.

Confidence at this stage comes from two places. One is evidence: you ran a small test and a few real people wanted the thing. The other is borrowed: someone who has been where you are looks at what you have and tells you, genuinely, that it is worth going for. I have watched that second one change what a founder does next more often than any framework ever has.

What to skip right now

A quick word on what to ignore for now. These all feel like progress, and at the idea stage they are mostly procrastination in a nicer outfit:

• Hunting for a technical co-founder. You can build the first version yourself now. Go find the co-founder once there is something real to co-own.
• Raising money. Investors fund evidence and momentum. At the idea stage you have neither yet, and that is completely fine.
• Applying to accelerators. Worth it later. Right now it pulls you away from talking to customers.
• Naming, logos, and the perfect domain. No company has ever failed because the logo was not ready.
• A detailed 12-month roadmap. You do not know enough yet for it to be anything but fiction.
• Building in secret. Almost nobody will steal your idea, and the feedback you give up by hiding costs you far more than the theft you are afraid of.

The sequence

None of this asks you to be technical, well-funded, or even all that confident on day one. It asks you to pick something you care about, get honest about whether anyone else does, build the cheapest version that teaches you something, and keep one person who has been there close enough to tell you when you are about to wander off.

I built GrowthMentor because hiring strangers off Upwork just to have someone to think with was a ridiculous solution to a real problem. Sixty thousand mentorship sessions later, the pattern I am most sure of is a simple one: the founders who move fastest are usually the ones who stopped trying to figure it all out by themselves.

So if you have an idea and a knot in your stomach about it, the knot is normal. Go say the sentence to someone who has the problem. See what happens to their face.

Founders who've launched from an idea

You've got the idea. There's a knot in your stomach,
talk it through with someone who's been here.

Bring your one-sentence problem, your rough prototype, and the worry that won't let you start. One call with a mentor who built from zero tells you whether to go for it or rethink.

Talk to a mentor