AI Information

Last updated: 19 August 2026

We use AI every day, in our own work and in yours. That is worth being open about, because if you are paying us to write and publish things under your name, you are entitled to know how they were made and what we check before they go live.

This page explains where AI is involved, which tools we use, what we will not let them do, and what happens to your information. If anything here is unclear, ask us and we will answer plainly.

How we use AI in your work

Writing

Pages, articles and answers are drafted with AI assistance and then edited by a person before anything is published. Every factual claim, number and date is checked against something real: your own information, a public register, or a source we can point you to. We do not publish anything under your name that has not been read by a human first.

Images

Some article images and illustrations are generated using AI tools, then processed to match a house style. Where we publish images of real work, they are real photographs supplied by the business they belong to. We do not present a generated image as a photograph of work that happened.

Social posts

Where we manage social posting, we turn photographs you send us into posts written to sound like you. The system is deliberately built so it cannot invent anything: if a photo arrives without a line from you describing the job, it refuses to write rather than making something up, and we come back and ask you instead. Every post is read by a person before it goes out.

Review replies

Where we manage reviews, replies are drafted with AI assistance in your business’s tone of voice, then reviewed and posted by a person. Nothing is ever published to your public profile automatically. We do not write reviews, and we would not work with anyone who asked us to.

Research and analysis

We use AI to help read and summarise large amounts of information: what people are searching for, what competitors are doing, what is on a website that should not be. The underlying data comes from measurement tools rather than from an AI model, because a model asked to estimate a number will produce a number whether or not it knows one.

How we use AI on this website

The assistant on this site

The chat assistant on our pages is powered by an AI model. It answers questions about what we do using a fixed set of information about our business. It cannot browse, it cannot see anything about you, and it is not able to quote a price. If you ask it something it does not know, it is instructed to say so and offer to put you in touch with a person.

If you leave your details through the assistant, a short summary of the conversation is sent to us by email so we can reply properly. That summary is generated by the same AI model.

Article images

Blog and guide header images are generated with an AI image tool and then colour processed to match our brand. They are illustrations, not photographs of anything.

Our own copy

Much of the writing on this site was drafted with AI assistance and edited by us. The claims, numbers and case study results are ours and are checked before publication. We do not publish a statistic we cannot evidence from our own work.

The tools we use

AI model requests are sent through OpenRouter, which routes them to the model provider. The models we currently use are:

• Claude (Anthropic) for writing, drafting and summarising
• Gemini (Google) for generating illustrations

We also use measurement and data services that are not AI models but do process information on our behalf, including search data providers, Google Search Console, Google Ads, our email provider and our hosting platform. These are listed in our privacy policy.

We choose tools on output quality and on how they treat data. We do not use AI services that train their models on what we send them.

What our AI is not allowed to do

Most of the work in using AI responsibly is deciding what it must never do, then building those limits in rather than hoping. Ours are enforced in code, not left to the model’s judgement:

• The assistant cannot quote prices. Its information contains no prices, and any figure that appears in a reply is stripped out before it reaches your screen. We test this deliberately, by trying to make it break its own rules.
• Nothing reaches a public profile without approval. Social posts and review replies can only be published after a person has read and approved them. That is a hard rule in the system, not a step someone could skip in a hurry.
• Nothing gets invented. Where we generate a post from your work, the tool is given only what you told us and is instructed never to add a number, a date, a guarantee or an award that you did not mention.
• No fabricated reviews, testimonials or results. Not now, not ever, at any price.

How we handle your information

What is sent to an AI service

When we draft something for you, the material we send is business information: what you do, where you work, notes and photographs you have given us, and your tone of voice. We do not send customer lists, personal data belonging to your customers, or anything about the people who visit your website.

Chat conversations

Messages you type into the assistant are sent to the AI provider to produce an answer. If you give us your details, the conversation summary comes to us. We do not use these conversations to train anything.

Your data is not training data

Nothing you give us is used to train an AI model, by us or by the providers we use. Your business information stays yours.

Where AI has been used on this site

For transparency, the following involved AI assistance:

• Page copy and service descriptions, drafted with AI and edited by us
• Blog articles and guides, drafted with AI and checked before publication
• Article header images, generated and then brand processed
• Answers in our FAQ sections, drafted with AI and verified against how we actually work
• The free tools, built with AI assistance, with the calculations based on real data rather than model estimates

Case study results, client names and any number presented as fact are not AI generated. They come from measured results and are used with permission.

Helping you get found by AI assistants

Worth separating two things that sound similar. Using AI to do work is what this page is about. Helping your business be the one an AI assistant recommends is a service we offer, and it is a different thing entirely.

When someone asks an assistant who to use for a job in your area, the answer comes from what those systems can find and trust about you. That work is about making your business clear, consistent and well evidenced online. There is no way to pay an AI assistant for a mention, and anyone who tells you otherwise is selling something that does not exist.

Our commitments

1. A person reads everything before it is published. Copy, posts, replies and images, without exception, whether it goes on our site or yours.
2. We tell you where AI was used. If you want to know how something we wrote for you was produced, ask and we will tell you straight.
3. We do not fabricate. No invented reviews, no made up results, no statistics we cannot show you the source for.
4. Your information is not training data. Not for us, and not for the providers we use.
5. We keep this page current. The tools change quickly. When what we use changes, this page changes with it.

Questions about how we use AI

If you would like to know more, or you want something about your own account explained, we are happy to go through it. Email hello@seopages.co.uk or call +44 203 051 6344.