You launched a website with an AI builder, but enquiries are not arriving. The tool alone does not explain why: traffic may be missing, the offer may not fit the audience, or the form may fail. Start with evidence before deciding to rebuild.
Treat the seven areas below as possible causes to investigate. For each, record the specific finding and missing data. The total alone does not establish why inquiries are low, how long repairs will take or whether to change platforms.
Seven possible causes
- Copy fits anyone, so it sells no one
- Proof the site offers no evidence that you are real
- Google does not know the site or does not understand it
- Placeholders leftover default text and the wrong language
- Measurement without it you are guessing
- Path to an inquiry it is hidden
- Platform ceiling limits exports and edits
1. The copy fits anyone, which means it sells no one
If a builder receives only a few generic words about the business, check whether its copy contains real information or phrases such as "Quality services with a personal approach." That phrase does not explain which jobs you take, where you work or what supports your offer. This is not a property of all AI copy; assess the actual brief and output.
How to spot it: Open your homepage and cover the logo. Could a competitor use the exact same text word for word? Is there no city, no specific services, no prices, no name? Does it say "our team of professionals" when it's just you?
What to do: Rewrite the key pages in your own words: what exactly you do, for whom, where, for how much, and what the visitor should do next. Keep the generated text as a skeleton if you like, just get yourself into it. How to use AI sensibly when building a site is something I've covered from the maker's side in AI and website building.
Using AI alone is not grounds for a penalty. Google's scaled content abuse policy concerns many pages created primarily to manipulate rankings rather than help people, regardless of the tool used. For ordinary business copy, check accuracy and usefulness; generic wording does not establish a penalty or a specific ranking loss.
Worked example: Same service, a more specific offer
ORIGINAL EXAMPLE · FICTIONAL DATA
Before
- "Quality services with a personal approach."
- "Pushing the boundaries of possibility."
After
- "We repair brakes and adjust your bike's gears."
- "Describe the problem. We will agree scope and timing."
Fictional workshop. No measured enquiry-growth claim.
2. There's not a single proof that you're real
An illustrative image can explain a subject, but it does not prove your premises or completed work. Distinguish illustrations from real photographs and check for accurate provider information and examples of actual work.
How to spot it: Count how many photos on the site are genuinely yours. Will a visitor find at least one testimonial with a full name? A photo of a finished job?
What to do: Add relevant photographs of real work, the brief and a supportable outcome. Publish testimonials and personal details with appropriate permission. A photo count alone does not guarantee trust or enquiries.
3. Google doesn't know your site, or doesn't understand it
Launching in a builder does not itself guarantee indexing. Check the settings, availability and content of important pages. A default "Home" title can prompt an improvement, but does not prove Google is unaware of the page.
How to spot it: Use site:yourcompany.com only as an initial check. Google warns that this list is incomplete. Verify a particular page with URL Inspection in Search Console. Separately check whether its title clearly describes its content.
What to do: Give important pages descriptive titles and check their status in Search Console. A sitemap helps discovery; it does not guarantee indexing. For a local service, also check your Google Business Profile.
4. Leftover placeholders and the wrong language
Default text can survive translation or template editing: "About us", "Submit" or somebody else's phone number. An incorrect contact directly obstructs communication; assess other defects while walking through the actual website.
How to spot it: Go through every single page, including the footer, the 404 page, the form confirmation, and the automatic email your customer receives. Look for leftover defaults, template text, and details that aren't yours.
What to do: List incorrect text and states, then check which are editable in the admin. Time depends on scope and access. After correcting them, have another person test the form and confirmation email.
5. You're not measuring, so you're guessing
Without measurement you can't tell apart two completely different situations: "nobody comes" and "people come but leave without inquiring." Each one calls for a different fix.
How to spot it: Open analytics and verify that visits and completed inquiries are actually recorded. Compare a test form submission with its delivered email. A missing or different count may indicate a measurement issue, not automatically zero demand.
What to do: Turn on analytics, builders usually have it built in, or connect a free analytics tool. Two numbers are enough to start: how many people came, and how many of them sent an inquiry. Pick the next step from those two numbers:
What to work on, by the numbers
- Nobody comes
- work on visibility (cause 3)
- They come but don't write
- work on copy (cause 1)
- work on proof (cause 2)
- work on the path to an inquiry (cause 6)
6. The path to an inquiry is hidden
I see it all the time: a nice website, but the phone number sits only in the footer, the form lives on a Contact subpage, and there's no button at all. A visitor who has decided to inquire has to hunt for your willingness to be contacted, and some give up along the way.
How to spot it: Open the site on your phone and time how long it takes you to make a call. Is the number tappable, or do you have to retype it? Is it clear on the first screen what you're supposed to do?
What to do: Put the phone number in the header as a tappable link, add a clear "Get a free quote" button right at the top, and cut the form down to name, contact, and message.
A prospect should easily understand how to continue. Steps depend on the task: a booking needs a time and a project may need supporting information. Test the entire journey, including errors and confirmation; no universal click count guarantees an inquiry.
Worked example: An interested visitor needs an obvious next step
ORIGINAL EXAMPLE · FICTIONAL DATA
Before
- Contact only in the footer.
- Fields: revenue, company ID, street, postcode, role, phone, email, message.
After
- Button: "Describe the bike repair".
- Fields: bike type, problem, contact.
- Confirmation: "Request received; we will use your supplied contact."
Illustrative confirmation; actual delivery needs a separate test.
7. You've hit the platform's ceiling
Export and editing options vary by platform and plan. Verify transfer of content, layout, forms, databases and integrations separately. A text export does not establish that the complete website can run elsewhere.
How to spot it: Test your site on mobile with PageSpeed Insights, try to find an "export site" feature in the admin, and ask yourself: if the platform raised its prices or shut down tomorrow, what exactly would you keep?
What to do: List the actual constraints and check exports, integrations, running costs and possible migration costs. If the platform supports the tasks you need, it does not need replacing even when the site generates work. Base a rebuild on a demonstrated constraint and total costs.
When a builder website is honestly enough
A simple builder website can suit an established business too. If its main role is to help referred customers verify the company, focus on accurate information, evidence of work, working contact paths and maintenance. Decide on further investment from its purpose and observed problems. I discuss small-budget scope and constraints in why a website for €200 is not enough.
How many causes fit your website?
- 1-2 findings: Verify cause and scope. A single broken form may matter more than several minor copy issues; the count does not determine repair time.
- 3-5 findings: Prioritise obstacles by impact and first repair the inquiry path. An audit establishes scope; the AI website fix offers an audit with a verdict for €200 and fixes from €600.
- 6-7 findings: Compare targeted repairs with a rebuild, including content migration and running costs. The total alone is no reason to leave the platform. A custom website from €800 is a separate offer, not the conclusion of this checklist.
What to take away
- The tool alone does not explain missing enquiries; start with evidence, not with a decision to rebuild the whole site.
- Treat the seven areas as possible causes to investigate, and record the specific finding and missing data for each.
- Without measurement you cannot tell "nobody comes" from "people leave without inquiring"; two numbers are enough to start: how many came and how many sent an inquiry.
- The phone number, a clear button and a short form should be visible at once; test the whole journey, including errors and confirmation.
- Decide on a rebuild or a platform change from a documented constraint and the total cost, not from the count of findings.
Choose the next step from the findings. Start with the website checklist or a free mini audit. An automated audit supports inspection; it does not measure why enquiries are missing. For the platform decision, see the AI website guide; paid diagnostic and repair scope is described under AI website repair.