Free tool

The thirty questions your customers ask, and the ones your site never answers

Paste your address. It reads your home page and a few of the pages linked from it, works out what you do, and returns the thirty questions a buyer asks before they get in touch. Each one drafted from your own copy, and flagged where your site does not answer it at all.

Read a site

Coming soon

This one is built and is not switched on yet. It goes live shortly, and when it does it will be free and ungated like the rest of them, with no sign-up and nothing stored.

In the meantime, the website test scores your home page against nine fixed checks, and it works now.

What it covers

What you do

The question a stranger asks first, and the one most home pages answer with a slogan instead of a sentence.

Coverage and eligibility

Do you come to my area, do you take my kind of job, do you work with someone like me.

How it works

What actually happens between getting in touch and the work being done.

Cost and quoting

Not a price. How cost gets decided, what it depends on, and whether a quote costs anything.

Timing

How soon, how long, and what your lead time looks like right now.

Trust and credentials

Licensing, insurance, how long you have been going, who has used you before.

Problems and guarantees

What happens if something goes wrong. Almost nobody publishes this and everybody wonders.

Getting started

The actual next step, in one instruction.

The gaps are the point

Almost every business we run this on gets the same shape of result. The questions about what they do and how to reach them are answered, often several times over. The questions about what happens when something goes wrong, how long things take right now, and how cost is decided are not answered anywhere.

That is not an accident. Those are the uncomfortable questions, and the instinct is to leave them for a conversation. The problem is that the buyer is deciding whether to have the conversation, and an unanswered question at that moment does not become a phone call. It becomes the next result in the search.

What to do with the list

Work the missing ones first, in the order the tool gives them. Write a real answer to each, in your own words, on a page a buyer will actually reach. Then do the half-answered ones, which are usually a sentence away from being fine.

The reason this matters more than it used to is that these questions are no longer only being read by people. They are the material an AI assistant uses when someone asks it to recommend a business like yours, and a site that does not answer a question cannot be quoted answering it.

If you are planning to put a chatbot on top of this, read chatbots that get fired first. A bot with nothing to answer from is the single most common way this goes wrong, and the FAQ work is the prerequisite rather than the thing the bot replaces.

For what a visitor judges before they even get to your FAQ, run the website test, and read the four second test.

The companion guide

Twelve reasons your AI gives generic answers, and what to do about each one

Where this applies

AI systems for firms whose most expensive people spend their day on admin

AI systems for studios and clinics where an empty slot is revenue that never comes back

AI systems for schools where a four week enrolment window decides the whole year

FAQ extractor FAQ

What does it actually read?
Your home page, and up to five more pages linked from it, chosen by relevance rather than by the order they appear. A page called shipping or FAQ is worth more than the sixth item in your navigation. It goes one level deep and no further, and it honours your robots.txt before requesting anything.
Where do the answers come from?
Your own copy, and only your own copy. Where a page answers a question, the draft is written from what that page says and the page it came from is named underneath. Where nothing on your site answers it, it says so rather than inventing something reasonable, which is the entire point of the exercise.
Why does it flag questions instead of just answering them?
Because the gaps are the useful output. A list of thirty answers you already have on your site is mildly interesting. A list of nine questions every buyer asks that your site never addresses is a work order. The filters default to putting those first.
Do you store the URL I enter?
Not in any durable sense, and it is worth being precise about what that means. There is no database and nothing is written to disk. A finished result is held in memory for about half an hour, keyed by a hash of the address rather than the address itself, so that running the same site twice does not pay for the analysis twice. That cache disappears whenever the server instance recycles, and it is the only thing that outlives your request.
It said there was no readable text on my site.
That usually means your content is drawn in by JavaScript after the page loads. This reads the document the way a first-time crawler sees it, which is also roughly how a search engine and an AI assistant see it. If your copy is invisible to this, that is worth knowing on its own.
Why is this one rate limited so much harder than your other tools?
Because it is the only one that fetches six pages and then does a substantial piece of analysis, so it costs real money every time somebody runs it. A few runs per half hour from the same place keeps it free and ungated for the people who actually need it.
Can I just paste these onto a FAQ page?
Edit them first. The drafts are accurate to what your site says, which is not the same as being right, and your site may be out of date. Read each one, fix what is wrong, and answer the gaps properly. The draft is meant to remove the blank page, not the thinking.
Does it work on a French site?
Yes. It writes the questions and the answers in the language the site is written in, so a French site returns French output without changing a setting.