Using AI for marketing without hurting your SEO

You can use AI to help with your marketing without damaging your search visibility, as long as a person who knows the subject checks every fact, adds what only your business knows and decides what gets published. Google does not rank pages lower because a machine helped write them. It ranks pages lower when they are thin, inaccurate or interchangeable with a hundred others, and unedited AI drafts are very good at being all three.

This post is about writing with AI: drafting web pages, blog posts, product copy and ads. If your question is how to get your business quoted inside ChatGPT, Perplexity or Google’s AI Overviews, that is a different job, covered in my guide to AI search optimisation.

What Google says about AI-written content

Google’s position has been stable since it published its guidance on AI-generated content on 8 February 2023: appropriate use of AI or automation is not against its guidelines, and what it rewards is original, helpful content however it was produced. What is against the rules is using automation mainly to manipulate rankings.

The rule that matters most in practice sits in Google’s spam policies (last updated 28 August 2026). Scaled content abuse means producing large numbers of pages to rank rather than to help anyone, and the first example Google gives is using generative AI tools to create many pages without adding value for users. The policy applies however the pages were made. A business that turns one service page into 200 near-identical town pages with an AI tool is squarely inside it.

Google’s separate page on using generative AI content on your website (last updated 1 October 2026) adds two points worth knowing. It reminds you that these models predict likely words rather than look up facts, so everything should be checked by a person before it goes live. And it says that check applies to page titles, meta descriptions, structured data and image alt text too, not only the body copy.

Why raw AI drafts rarely earn a place in search

A generative AI tool writes the most probable answer to your prompt, which is roughly the average of what has already been published on the topic. That is useful as a starting point and weak as a finished page, for reasons that are easy to spot once you know them.

• It adds nothing new. If your page says what the top ten results already say, Google has no reason to show yours instead of theirs.
• It states things confidently that are not true. A hallucination can be an invented statistic, a regulation that does not exist or a quote nobody said. On a business site, that is a liability as well as an SEO problem.
• It has no experience to draw on. It has never fitted a boiler in a Victorian terrace or handled a lease extension in Kensington. The detail that shows real knowledge is exactly the part it cannot supply.
• It drifts into American defaults. A z where British English uses an s, prices in dollars, sales tax instead of VAT, attorneys instead of solicitors, and US law presented as if it applied here. UK readers notice, and it undermines trust in everything else on the page.
• It answers a slightly different question. Without a clear brief, a draft often wanders between informational and commercial intent, so the page satisfies neither type of searcher.

Google’s quality guidance describes the qualities it looks for under E-E-A-T: experience, expertise, authoritativeness and trust. Every item in the list above cuts against at least one of them.

Where AI is genuinely useful in a small marketing team

The work AI tools do well is the work that sits around the writing, not the judgement at its centre.

• Summarising long source material, such as a regulator’s guidance or your own sales call notes, so you can see what is worth reading in full.
• Suggesting angles and questions you might have missed when planning a page, which you then check against what customers actually ask.
• Turning your notes into a rough first draft that you rewrite, which is faster than starting from a blank page.
• Producing several versions of an ad headline or meta description to choose from and edit.
• Reworking existing content for another format, such as turning a guide into an email or a set of social posts.
• Grouping a long keyword list or a pile of customer reviews into themes before you look at them properly.

In each case the tool saves time on a task where you can see at a glance whether the output is right. That is the test I would apply to any new use: if you cannot judge the output quickly, the tool is not saving you time.

The rules I would hold any AI-assisted content to

Do AI detectors matter for SEO?

I would not spend money or time on them. Detection tools guess, they disagree with each other, and they regularly flag text people wrote themselves. More to the point, Google has said it judges content on quality rather than on how it was produced, so passing a detector proves nothing about whether a page will rank. Rewriting a good page to fool a detector usually makes it worse. Put that effort into making the page more accurate and more specific.

A workflow that keeps AI in its place

1. Start with a brief, written by a person. Decide the search term, the reader, the one question the page answers and the action you want them to take. My SEO content brief template gives you the fields to fill in.
2. Gather the material only you have. Prices, process, examples, photos, common objections, your opinion on the question. Write it down in rough notes.
3. Use AI for the first draft if it helps, giving it your brief and notes rather than a one-line prompt, and asking for British English.
4. Rewrite in your own voice. Cut generic paragraphs, add specifics, and make sure the page answers the question in its opening lines.
5. Check every fact against a primary source and remove anything you cannot verify.
6. Write or check the metadata by hand. Title, meta description, image alt text and any structured data all appear in search results, so they need the same review as the body.
7. Read it on a phone as a customer would, then publish, and add internal links to and from related pages.

For a service page in particular, my post on how to write a service page that ranks and converts covers structure and wording in more detail.

Signs AI content is already holding your site back

If a lot of AI-assisted content has gone live without much review, these are the patterns I look for in Google Search Console and on the site itself:

• Many pages discovered or crawled but not indexed, which can mean Google does not see enough value in them to store them.
• Several pages taking turns to appear for the same search, none of them holding a steady position.
• Blog posts that read alike, open the same way and could sit on any competitor’s site unchanged.
• Pages with impressions but almost no clicks, often because the title and description are generic.
• Facts that have quietly gone out of date, or were never right.

The fix is rarely to delete everything. It is usually to merge overlapping pages, improve the ones that earn impressions, and remove or redirect the rest. That is the kind of work covered by on-page SEO, and if you need a plan for what to publish next and why, content SEO and strategy starts there. If the bigger question is how your business appears in AI-generated answers, my AI search optimisation service looks at that separately.

The short version

AI is a capable assistant and a poor author. Use it to save time on research, structure and first drafts. Keep the facts, the experience and the final decision with a person. Publish only what you would be happy to put your name to, at a pace you can check properly. Do that and Google has no reason to treat your pages differently from anything else you write. If you would like a second opinion on content already on your site, send it to me and I will tell you what I would fix first.