AI SEO — what AI changes in SEO work, and what stays human

AI SEO means two things at once, and they get mixed up constantly: AI doing a growing share of the SEO work itself — research, drafting, internal linking, technical monitoring — and search itself becoming AI-mediated, so your content has to work as a source for language models. The first changes how an agency works. The second changes what the work is supposed to achieve. Both are happening now, and both can absolutely be done badly. Here's an honest look at what AI actually contributes to SEO today, what still requires people — and which questions to ask an agency that calls itself an "AI SEO agency".

The short version

• AI SEO = AI as a tool in SEO work and SEO for AI-driven search surfaces. Two different things.
• AI is good at volume and monitoring: research, drafts, internal linking, technical review.
• Humans still have to own strategy, facts, prioritisation and accountability. AI has none of these.
• A serious AI SEO agency can answer concretely what the AI does and who checks the result.
• Guarantees of rankings or citations are a red flag — no matter how much AI is involved.

AI SEO means two things at the same time

When someone says "AI SEO", they mean either the tool or the goal — and the difference isn't pedantry. AI as a tool means parts of the workflow get automated, so more of your budget goes to judgement instead of routine. SEO for AI means a growing share of searches are answered by ChatGPT, Gemini, Perplexity and Copilot, and visibility means being cited there — not just ranked on Google. We've covered that second part in our article on how language models pick their sources. This article is mostly about the first: what's happening to the craft itself.

What AI actually does in SEO work today

Used properly, AI is an excellent assistant for the parts of SEO that are volume and systematics:

• Research. Grouping thousands of search phrases by intent, finding gaps between your content and what people actually ask, summarising what ranks and why. What took days now takes hours.
• Drafting. First versions of texts, meta descriptions and FAQs. Drafts — not finished content. The difference is roughly the one between a sketch and a house.
• Internal linking. Suggesting which pages should point to each other, and catching orphaned pages nobody links to.
• Technical monitoring. Continuous review of broken links, indexing errors, slow pages and pages that suddenly drop — instead of a manual check once a quarter.

The net effect isn't that SEO becomes cheaper to do badly (it already was), but that an agency can spend more of your time on the things that move numbers.

What still has to be human

Just as important is the list of what AI cannot own:

• Strategy and prioritisation. AI can list a hundred opportunities. It doesn't know which three matter for your specific business, market and margins.
• Facts and accountability. Language models write convincingly even when they're wrong. Prices, claims and expert content must be checked by someone who can stand behind them afterwards. That's a legal and reputational reality, not nostalgia.
• Quality and voice. Mass-produced AI text has a smell, and both readers and search engines have developed a nose for it. Text that's supposed to build trust needs a human at the end of the line.
• Relationships and authority. Mentions, partnerships and genuine reviews — the signals that actually build credibility — can't be automated into existence. This ties straight into the broader content and marketing work.

The questions to ask an "AI SEO agency"

"AI" in a sales pitch says nothing about quality — it can mean a smart workflow or mass-produced filler. Ask concretely:

1. What exactly does the AI do at your agency — and what do the humans do? Vague answers here are answers.
2. Who fact-checks the content before it's published? If nobody, you know what you're paying for.
3. What do you report on? Enquiries and revenue, or just positions and undefined "AI visibility"?
4. Do you also work on visibility in AI search — and how? Ask for mechanisms (structure, schema, llms.txt, mentions), not buzzwords.
5. Do you guarantee rankings or citations? The correct answer is no. Nobody controls what Google ranks or what a language model cites, and anyone promising it is selling something they don't own.

What does AI SEO cost?

With us, AI SEO isn't a separate, more expensive shelf — it's how we work. Ongoing SEO and AI search work starts at NOK 5,500/month, and pure local SEO at NOK 8,700/year. AI means more of that time goes to judgement and content that works, less to routine. The details are on our pricing page.

The future of SEO is still SEO

The honest answer about the future of SEO is that the foundation isn't moving: technical accessibility, content that answers real questions, and credibility others can see. AI changes the pace of the work and adds new surfaces to be visible on — as of today especially Perplexity, Gemini and Copilot alongside ChatGPT — but it doesn't change the fact that quality wins over time. If you want to know what this looks like for your business, book a short call, or read more about how we work with AI search and SEO.