AI is changing SEO in two ways
First, the work of SEO is changing: time-consuming steps such as research, analysis and drafting can be done much faster with AI. Second, search itself is changing: Google’s AI summaries and tools like ChatGPT and Perplexity give users direct answers instead of a list of links. We cover that second side, getting your brand into AI answers, in our GEO (Generative Engine Optimization) guide. This article focuses on the first: how to use AI in your SEO work.
What AI genuinely speeds up
- Keyword grouping: clustering hundreds of queries by search intent (informational, comparison, purchase) and planning which page targets which group.
- Content briefs: listing the questions users ask about a topic, missing subtopics and the points competing pages cover.
- Meta title and description drafts: alternatives that fit character limits and use the keyword naturally.
- Structured data: first drafts of schema markup such as Organization, Article and FAQ.
- Technical analysis: summarising crawl reports and Search Console data, and picking out recurring issues and priorities.
- Internal linking suggestions: proposing which pages should link to which, based on content similarity.
- Multilingual content: drafting translations, with the final read left to a native-speaking editor.
What you should not hand over to AI
- Strategy and positioning: which audience you speak to, with which service and what difference is a business decision; a model cannot know it for you.
- Facts and claims: prices, statistics, legal information and technical details can be invented by AI. Everything you publish should rest on a source.
- Experience and expertise: knowledge from real projects, client questions and the field is what sets content apart. AI cannot produce it, only help organise it.
- Final editing: a person should always read before publishing, for brand voice, accuracy and readability.
- Mass page generation: auto-publishing hundreds of near-identical pages looks fast in the short term but creates serious risk.
What does Google say about AI-written content?
Google’s approach looks at how helpful content is, not how it was produced: content written for people, reliable and genuinely answering a question is valuable even if AI helped make it. On the other hand, low-value content produced at scale primarily to manipulate rankings falls under its spam policies. In short, the question is not “did we use AI?” but “does this page actually help the reader?”
AI speeds up the draft; your knowledge and experience are what make the content worth reading.
A practical workflow
- Start with data: Search Console queries, keyword volumes and the questions your clients actually ask.
- Group and prioritise with AI: which topics need a new page and which need an existing page strengthened.
- Build the brief together: let AI pull out the questions and subheadings, and add your own examples and point of difference.
- Write the draft and verify it: check every number, claim and technical detail against its source.
- Complete the technical layer: title, description, schema, internal links and image alt text.
- Measure and update: after a few weeks, look at impressions, clicks and rankings in Search Console and improve the content based on real data.
What to ask when choosing an AI or SEO agency
- At which steps do you use AI, and who reviews content before it goes live?
- How do you ensure accuracy, and where does information about our brand come from?
- How do you measure success: rankings alone, or traffic, enquiries and conversions?
- Who keeps our data and accounts (Search Console, Analytics, content)?
- Do they guarantee rankings? No agency can guarantee a specific position on Google; a promise like that is a warning sign.
Common mistakes
- Publishing AI output unedited: the content stays generic, repetitive and unbranded.
- Having AI rewrite competitor content: a page that adds nothing new cannot beat the results already there.
- Skipping the technical foundation: on a slow, uncrawlable or mislabelled site, even good content stays invisible.
- Not measuring: producing more without knowing which pages work.
Conclusion
AI cuts research and production time in SEO considerably, but it does not supply strategy, accuracy or experience for you. The best results come from combining AI’s speed with human editing, real data and a solid technical foundation.
At SOART we run SEO and GEO work with AI-assisted tools, with every step reviewed by a person. If you would like to talk about your site’s search and AI visibility, get in touch.