The ChatGPT Prompt That Plans Your Website and Writes the Brief

TL;DR. Paste the interviewer prompt below into ChatGPT, Claude, or Gemini and answer its questions for about ten minutes. You get a one-page website brief, seven fields, under 300 words, that any builder, freelancer, or agency can quote from. Run the reality-check prompt before spending anything. And if you want it built by humans, our $399 Overnight Site takes the finished brief as direct input and goes live in a day.

The prompt: paste this into ChatGPT, Claude, or Gemini

Copy the block below into a fresh chat. It runs the same in ChatGPT, Claude, and Gemini. Answer its questions one at a time, honestly, and in about ten minutes you'll have a one-page brief that any freelancer, agency, or studio can turn into a real quote.

You are a senior web developer who scopes small-business websites for a living. Interview me about the website I need, then produce a one-page website brief that any builder, freelancer, or agency can quote from.

Rules for the interview:
- Ask ONE question at a time and wait for my answer. No multi-part questions.
- Ask at most 8 main questions, plus a follow-up when an answer is vague. Follow-ups don't count against the 8. Skip any question you can already answer from what I've said.
- Push back on soft answers. "Everyone" is not a customer. "Get more business" is not a goal.
- If I give the site more than one job, make me pick one. A site with three jobs does none of them.

Cover, in this order: what my business does and for whom, in one sentence; the ONE job of the site, meaning the single action a visitor should take (call, book, buy, fill a form); which pages the site needs, challenging any page that has no job; what copy and photos I already have versus what still needs making; whether I own a domain and whether anything is hosted today; what I'll honestly spend; and when the site must be live, with the reason behind the date.

Then output "WEBSITE BRIEF" with exactly these seven fields:
1. Business: what it does and for whom, one sentence
2. The one job of the site: the single conversion, written as a visitor action
3. Pages: the list, with one line on what each page must do
4. Assets on hand: the copy and photos I have, and what's missing
5. Domain and hosting: what exists today
6. Budget band
7. Deadline, and why that date

Keep the whole brief under 300 words. Plain language only. Include nothing I didn't confirm, invent no numbers, and write "not decided" wherever I haven't decided.

Copy this prompt

Two habits make the output better. Answer with specifics, because the model mirrors whatever clarity you give it: vague in, confident padding out. And when it challenges your page list, let it win. Most small-business sites need fewer pages than their owners think, and cutting them is the biggest lever on what the site will cost.

The seven questions to answer before building a website

The prompt outputs seven fields because those are the seven questions every builder asks before quoting, and the ones most owners show up without.

1. What the business does and for whom. One sentence. If this takes a paragraph, the site will be confusing too.
2. The one job of the site. The single action a visitor should take: call, book, buy, or fill a form. Pick one. Everything on the site either serves it or gets cut.
3. The page list. Every page needs a job. Pages that exist out of habit are cost without purpose.
4. Copy and photos on hand. Missing assets are what turns a two-week website into a two-month one.
5. Domain and hosting status. Thirty seconds to answer, and it changes the quote.
6. Budget band. Not a trap. A builder who knows your band proposes something that fits it instead of guessing high.
7. Deadline, and why. A date with a reason behind it gets treated seriously. "ASAP" does not.

"Not decided" is a fine answer for any of them. It's honest, and it marks where your first builder conversation starts. Walking in with these seven answered puts you ahead of most owners looking to hire someone to build a website, who typically show up with a Pinterest board and a feeling. And if you're not sure the thing you need is a website at all, settle website or app first before you brief anyone.

Why one interview beats a list of 150 prompts

Search this topic and you'll find pages listing 15, 30, or 155 prompts for planning a website. None of them end in an artifact. You read them, nod, and still have nothing to hand a developer.

The one-shot version is worse. Ask an AI to write you a website plan or a website brief in one go and it will, instantly and confidently, invent your customers, your pages, and your budget, because it doesn't know any of them. You get a document full of decisions nobody made, and builders quote defensively for those.

The interview flips the division of labor. The model asks and pushes back, which is what it's genuinely good at. You decide, which only you can do. Every line in the finished brief is something you confirmed out loud, and that's the whole difference between a brief that gets a real quote and one that gets a discovery phase. It's the same trick behind our app requirements prompt, which does this for founders scoping software.

The reality-check prompt: will this website do anything?

Now the uncomfortable one. Small-business websites rarely fail from bad design. They fail because nobody visits, and nobody planned for anyone to visit. A website is a destination, not a traffic source. This prompt makes your AI hold that line before you spend a dollar.

Here is my website brief: [paste your brief]

Assume this website gets zero visitors and zero conversions until my answers prove otherwise, and say "THIS SITE WILL DO ITS JOB" only on specific evidence, never on my enthusiasm.

Ask me these three questions, one at a time, and wait for each answer:
1. Where will visitors come from? Name the source: specific Google searches, a Google Business Profile, social posts, paid ads, an email list, referrals, a link in my bio. "People will find it" is not a source.
2. What is the ONE thing a visitor should do on the site, and how will I know each time it happens?
3. What will I do every week to send people to the site? Name the action and the hours it takes.

Flag every vague answer and make me replace it with something specific before you move on. Then give a one-paragraph verdict: THIS SITE WILL DO ITS JOB, or THIS SITE WILL SIT THERE. Whatever the verdict, name the single cheapest change to my plan (not to the site) that would most improve it.

Copy this prompt

Run it in a fresh chat so the model isn't primed to be agreeable about a brief it helped you write. If question one stumps you, that's the finding, not a failure. One source most owners never check is the AI engines themselves: whether ChatGPT recommends your business takes five minutes to test. And if the verdict says the plan itself is unproven, a one-page landing page is the cheapest way to find out whether anyone cares before you build the full site.

The fit-check prompt: who should build it?

Brief in hand, the last question is who builds the thing. Paste the brief into this and your AI scores the four realistic routes against your actual budget and deadline, instead of the generic advice every listicle recycles.

Here is my finished website brief: [paste your brief]

Score four ways of getting this site built against MY stated budget band and deadline: a DIY website builder (including AI builders), a freelancer, a traditional agency, and a fixed-price studio.

Output a table first: one row per option, with columns for typical cost, typical timeline, fit with my budget band, fit with my deadline, and the biggest risk for my specific brief. Wherever my brief is silent on something you need, write "not stated" instead of guessing.

After the table, one paragraph: which option fits this brief best and why, and the one question I should settle before spending money. If two options genuinely tie, say so rather than forcing a winner.

Copy this prompt

Whatever it recommends, hold every route to the same bar: a real number before work starts, and a site that doesn't look like a template. Cheap and good aren't opposites, but they only overlap when someone senior makes the calls. Cheap website design that doesn't look cheap shows what that looks like in practice.

Want ChatGPT to actually build the site? Sometimes you should

Some of you didn't come here for a brief. You want the AI to build the actual site, and honestly, sometimes it should. AI site builders produce a workable DIY site in an afternoon now. If your budget is zero, the site is for a hobby, or you're testing whether anyone cares about an idea, build it yourself and skip the rest of this section. That's the right call, and no developer should tell you otherwise.

The line is conversion. The moment the site's job is to make the phone ring, take bookings, or turn visitors into paying customers for a business you live on, the DIY ceiling shows up fast: template look, slow pages, forms that go nowhere, and nobody accountable when it breaks at 9pm on a Friday. We wrote the honest breakdown at I built my website with AI, do I need a developer if you're already down that road.

Here's what makes the brief worth ten minutes either way: it's a direct input to the human route. Our $399 Overnight Site takes the brief exactly as the prompt outputs it and puts the site live in a day. Here's how that works. Same artifact, either path.

What to do with the finished brief

Three honest routes out.

Build it yourself. If your budget band came out at zero, or the reality check said the plan needs proof before the site deserves money, DIY it. The brief still earns its keep: it stops you building six pages when one would do.

Send it to two or three builders. The brief is builder-neutral by design. Ask each for a fixed price against it and watch what they do with the budget band. Anyone who quotes without reading it has told you everything. Hire someone to build a website compares every route, with prices.

Or hand it to us. The $399 Overnight Site takes the brief as direct input and goes live in a day. If the one job of your site involves capturing leads and following up, the Engine starts at $699 and adds the pipeline: forms to CRM, booking, follow-up. Custom builds are $1,500 when the brief outgrows both. Every engagement is a fixed price agreed before we start, and you own the code from day one.

One last reason we give these prompts away: more buyers now ask their AI before they ask Google. A recent DappaSol lead found us by asking Gemini for help with their problem. The brief you just wrote is you doing the same thing, ten minutes ahead of everyone else.