AI Accuracy & Limitations

Sitemo uses artificial intelligence to generate and edit websites from what you describe. This page sets out what those features can do, what they get wrong, and what remains your responsibility before anything is published.

It forms part of the Terms of Service. How your prompts and content are handled when they reach an AI provider is described in the Privacy Policy.

AI is involved when you:

• generate a site from a description;
• ask for a change in chat, and the change is applied for you;
• have copy written or rewritten for a page;
• generate an image, or have one suggested.

Everything else — the editor, publishing, forms, your data — is ordinary software and behaves predictably. The limits below apply to the generated parts.

Generative models produce text and images that are statistically plausible, which is not the same as correct. In practice that means:

• Invented facts. Opening hours, prices, years in business, qualifications, awards, service areas — a model will fill a gap with something believable rather than leave it blank.
• Confident tone regardless of accuracy. Output doesn't signal when it is guessing, so wrong statements read exactly like right ones.
• Out-of-date knowledge. A model's training has a cut-off date and it has no knowledge of your business beyond what you tell it.
• Generic or repeated results. Two businesses described similarly can receive similar copy and layouts. We don't guarantee that output is unique.
• Imperfect images. Generated images can contain distorted details, unreadable text, or objects that don't exist as depicted.
• Inconsistency. The same request can produce different results on different runs; output is not reproducible.
• Bias. Models reflect patterns in their training data, which can carry cultural and other biases into wording and imagery.
• Language and locale. Quality varies by language, and conventions like dates, currency, addresses and spelling may not match your market.

Treat generated output as a first draft. Before publishing, check at least the things a visitor could act on and be misled by:

• contact details, address, opening hours and service areas;
• prices, packages, and anything that sounds like an offer or a guarantee;
• claims about experience, qualifications, certifications, awards or memberships;
• names of people, products, partners and brands;
• any statistic, date or number.

You are responsible for everything published on your site, whoever or whatever drafted it. Publishing is your decision and your confirmation that you've reviewed the content.

Generated output is not checked against advertising rules, professional codes of conduct, or sector-specific regulation. If your work is regulated — health, medical, legal, financial, insurance, food, alcohol, gambling, childcare, property, recruitment and others — the wording on your site must be reviewed against the rules that apply to you before it goes live.

Nothing generated by the service is legal, financial, medical or other professional advice, to you or to your visitors. Do not rely on it as such, and don't publish it in a way that invites your visitors to.

The same applies to legal text on your own site: privacy notices, terms, cookie notices and accessibility statements need to reflect what your business actually does.

Generated output may resemble existing material, and we can't warrant that it is original or free of third-party rights. Before publishing, take care that names, slogans, logos and images don't conflict with someone else's trademark or copyright.

As between you and us, output generated for your site is yours to use, subject to the Terms of Service. Whether generated material attracts copyright protection at all varies by country, so treat it as unprotected unless you've taken advice.

AI features are provided on an "as is" basis. We don't guarantee that a generation will succeed, that it will match what you had in mind, that a particular style or structure can be reproduced, or that output will rank, convert or perform in any particular way.

The underlying models and providers change over time as they are improved or replaced. Results may differ from one period to another, and a request that worked before may produce something different later.

Your prompts and the content being worked on are sent to third-party AI providers so they can be processed. Under our agreements with them, that content is not used to train their models, and we don't train models on it either — see the Privacy Policy.

Please don't put anything into a prompt that shouldn't be processed by a third party: passwords, card numbers, government identifiers, health records, or other people's confidential information.

If a generation produces something inaccurate, offensive, or that appears to infringe someone's rights, please tell us at support@sitemo.ai or through our contact page. Reports help us improve the prompts and safeguards behind these features.