AI Website Building
Build and sell websites in an afternoon instead of a fortnight.
Build queue
Brief to live site to invoice
- Dental practice · 4 pages78%Copy and structure
- Roofing firm · 6 pages34%Generating pages
| Event | Project | Invoice |
|---|---|---|
| Brief in | Roofing firm, 6 pages | — |
| Care plan started | Monthly upkeep | +€95 |
| Invoiced | Dental practice, 4 pages | +€1,450 |
| Live | praktijkhavenzicht.nl | — |
| Human pass | Copy and structure | — |
Example figures. These amounts show how the workflow behaves — they are not earnings, not ours and not anyone else's. Nobody made this money.
Open full screen · recording mode — sized for vertical video.
Follow these prompts to start building your income stream
7 steps. Paste each prompt into Claude, use what comes out, move to the next. The last one ends in something live — not in a plan to make something live.
- 01
Pick your trade and your price
Free sampleOne trade, one fixed price, one page of text you can send today.
In this step you stop being a web builder for everyone and become the web builder for one trade. That single decision is what makes the rest of the plan fast: the offer writes itself, the outreach writes itself, and the second client costs a fraction of the first. You end this step with the text of your offer, ready to put online.
1 — Pick the tradeI am starting a productised website service and I need to pick one trade to sell to. My city or region: [REGION] Trades I already have access to, through work, family or hobbies: [YOUR ACCESS] For each trade I listed, give me: 1. The five questions a customer asks before they call, in the customer's own words. 2. What their current websites almost always get wrong. 3. What the owner cares about, phrased the way that owner would say it. 4. One reason this trade might be a bad first choice. Then rank them by how easily someone with no portfolio could get a paying client this month, and tell me in one paragraph what makes the top one easier than the rest.
2 — Write the offerWrite the offer page for a productised website service for [TRADE] businesses. Build tool I use: [BUILD TOOL] Honest delivery time, including my day job: [DELIVERY TIME] Fixed build price: [BUILD PRICE] Monthly maintenance price: [MAINTENANCE PRICE] Write, in this order: - One sentence on what the client ends up with. Their outcome, not my process. - What is included, as a list a non-technical owner can tick off. - What is not included, so scope creep has somewhere to stop. - The steps from yes to live, with who does what. - What I need from them, and what happens if it arrives late. - The maintenance plan as the obvious next step, not an upsell. Rules: no hourly rate anywhere. No mention of AI — they are buying a website. No superlatives. Plain sentences a busy person reads in ninety seconds. Keep the whole thing under 400 words.
Claude. Ask it to keep the offer text in a separate block so you can copy it straight out.
- [REGION]
- The area you can reach for a coffee. Local beats national when you have no portfolio.
- [YOUR ACCESS]
- Trades where you know someone or understand the jargon. Access beats market size at this stage.
- [TRADE]
- The one you picked with the first prompt. Fill it in before running the second.
- [BUILD TOOL]
- Lovable, Framer, Webflow — whichever you actually know.
- [DELIVERY TIME]
- The honest number. Late delivery kills referrals faster than a high price.
- [BUILD PRICE]
- Fixed. Your speed is the product; hourly billing charges you for being fast.
- [MAINTENANCE PRICE]
- Monthly. This is what makes the business worth owning in year two.
Then: Read the offer out loud to someone in that trade. Every line they answer with 'so what does that mean' is still marketing language — paste their reaction back into Claude and ask for a plainer version.
- 02
Put your own page online
In the packageA live one-page site with your offer, your price and a contact form.
In this step your offer stops being a document and becomes a link. You need it before you talk to anyone: the first question after 'what do you do' is 'do you have something I can look at'. One page is enough — you are selling websites, so this page is also your portfolio piece.
- 03
Line up ten businesses to call
In the packageTen named businesses and the message you send each one.
In this step you go from having an offer to having a pipeline. Ten is deliberate: enough to learn what people say back, small enough to do in an evening. You are not selling in the message — you are trying to get one reply.
- 04
Get their details, write their text
In the packageA filled-in intake and the full text of their site, in their words.
In this step someone said yes and you start delivering. The intake is what keeps a fixed-price build profitable: everything you need in one pass, so you are not chasing answers for three weeks. The copy comes straight out of their answers, which is why the intake questions matter more than the writing.
- 05
Build the site and check it
In the packageA finished site that passes a real check before handover.
In this step the site gets built and, more importantly, checked. The check is what separates you from the cheapest competitor: generated sites usually work on the builder's laptop and break on a customer's phone. You end with something you are willing to put your name on.
- 06
Hand over and invoice
In the packageThe site live on their domain, invoice sent, monthly plan agreed.
In this step you get paid and set up the part that makes this a business instead of a series of one-off jobs. Handover is a moment where clients quietly panic about being locked in, so the document you send matters as much as the site.
- 07
Turn one client into three
In the packageA case study you can send, and two referrals to approach.
In this step the first job starts paying for the second. A finished site and a happy owner are the only proof you need, and this is the week to ask — enthusiasm fades fast. You end with something to show and someone to show it to.
6 of the 7 steps are in AI Money
The sample above is a real step, not a teaser — it is the same depth as the rest. The package adds the other 6 for this strategy, and the same treatment for the other eleven.
Your first 7 days
- 1Build two example sites for businesses you know, from a real brief, and get them online with real copy.
- 2Write one page describing exactly what a client gets, in what time, for what price.
- 3List 20 local businesses whose current site is slow, unclear or missing.
- 4Send 10 of them a short message with a link to the closest example you built.
Your first 30 days
- 1Deliver the first paid site and record how many hours it actually took.
- 2Turn the intake conversation into a fixed questionnaire so you stop reinventing it.
- 3Add a maintenance plan to the offer and sell it to the first client before launch.
- 4Ask every delivered client for one referral inside their own trade.
What you need
core means you cannot skip it. The rest only if you need it.
- Lovablecorefree planGenerate the site from a brief
- Framercorefree planDesign-led alternative with hosting included
- Webflowcorefree planWhere clients need a CMS they can edit
- v0 (Vercel)free planComponent and layout generation
- Claudefree planCopy drafting and intake summaries
- Canvafree planClient-ready imagery without a designer
- 10WebWordPress route for clients already on WordPress
Want the research?— what it is, what it costs, and how we scored it
What you can actually watch
A brief going in, pages appearing, a live URL coming out.
What it is
You build websites for small businesses using AI site builders, and you charge a fixed price per site instead of an hourly rate. The generator does in an afternoon what used to take a fortnight; you do the part it cannot do — deciding what the site should say, to whom, and in what order.
Why this opportunity
This is the strongest of the twelve, and it earns that on evidence rather than enthusiasm. Published prices exist on both sides of the trade, which is rare in this field: the tooling is $25–$50 a month (Lovable, Tier 1, lovable.dev) while Webflow freelance rates run $25–$200+ an hour (Tier 3, webflow.jobs). You are not inventing a market — you are entering one that already pays, with a cost base that collapsed.
Who this is for
For someone who can talk to a local business owner without using the word AI, and who has the patience to write copy that says something. It is a poor fit if you want to avoid selling: the bottleneck here is finding the third client, never building the third site.
The business model
Productised: one fixed price for a defined deliverable, optionally with a monthly hosting and maintenance plan underneath it. Brennan Dunn's 2025 freelancer survey found that freelancers selling a productised service earn roughly 45% more per project than those billing hourly (Tier 3, LIKELY — primary source still to be verified).
How money is made
A one-off build fee per site, plus recurring hosting, edits and maintenance. The build fee pays this month; the maintenance plan is what makes the business worth owning in year two.
Who pays
Small and medium businesses, freelancers, local service providers, and startups that have not had a design round yet. The buyer is usually the owner, which means one conversation and one decision instead of a committee.
Finding customers
Local and sector-specific. This is one of the few AI services you can sell in your own city without explaining what AI is — the client buys a website, not a technology. Referral inside one trade travels fast; generic cold outreach to "businesses" does not.
Delivery
Intake and content gathering, then an AI-generated concept, then human editing, then launch, then hosting and maintenance. The editing pass is the deliverable; a generated site handed over unedited is what your competitors sell.
Where AI actually helps
Very high, and concentrated in one place: build time has dropped by an order of magnitude. It has not dropped the time needed for intake, copy decisions or client management — which is exactly why the work still pays.
Skills you need
Low to medium to start, medium to be good. You need taste, plain writing, and enough technical care to connect a domain and a form. You do not need to code.
What it costs to start
Very low. Lovable Pro is $25 a month, Business from $50 a month (Tier 1, lovable.dev, LAST VERIFIED: 2026-09-07); a domain and hosting on top. No stock, no staff, no advance.
How to price it
Fixed price per site, with tiers by page count and complexity, plus an optional monthly plan. Do not publish an hourly rate: your speed is the product, and hourly pricing charges you for being fast.
Common mistakes
Selling "a website" instead of an outcome for one kind of business. Handing over generated copy nobody edited. And the one with teeth: adding forms, logins or payments without a security pass — AI-generated applications demonstrably ship with vulnerabilities, and that moves you from brochure-site risk into the risk profile of AI micro-SaaS.
Scaling
Scalability scored 6 of 10 — real but bounded. Growth comes from repeatable niches (one trade, one template, one content pattern) and from the maintenance layer, not from taking on every kind of site.
The version that holds up
Pick one trade, build a template that already knows that trade's objections, and sell the same site repeatedly with the copy as the only variable. At that point you are selling a system, and the European Accessibility Act (in force since 28 June 2025 for much of e-commerce; LIKELY, to be verified) becomes a selling point rather than a chore.
Time to first revenue
1–4 weeks
1–4 weeks to the first paid site, assuming you sell locally or inside your own network and have a portfolio of two or three examples you made yourself. Without any selling effort, weeks is not realistic — the bottleneck is acquisition, not production.
How it scored
Ten axes, scored. Risk scored 5 out of 10, where higher is worse.
The state of this market
The best-evidenced opportunity of the twelve. The catch is commoditisation: everyone can generate a site now, so competition scored 8 out of 10. The money is in knowing what to build and for whom, not in the generating.
Not sure this one fits you?
Answer a few questions about budget, time and skills and see which of the eighteen match your situation — and which to avoid.
Run the Opportunity Finder