AI SEO consultant: what the label actually adds

An AI SEO consultant is an SEO consultant who also works on whether your business gets named inside AI-written answers: Google’s AI Overviews and AI Mode, and assistants such as ChatGPT. That second part is real work. It is not a separate technique, and Google says so in writing: its guide for site owners states there are no additional requirements and no special optimisations needed to appear in AI Overviews or AI Mode. What the label should add is measurement and a different kind of publishing. If it adds only a higher price for the same audit, you are paying for the word.

I searched “ai seo consultant” in Canada on 6 October 2026 to see who is selling it and what they promise. Here is what came back, what Google’s own documentation says, and the four claims I would test before hiring anybody under that title, me included.

Who is selling it, as recorded on 6 October

The page opened with an AI Overview, which is a fair joke on the query. Below it were organic results, a people also ask box, and a block of forum threads.

Three of the organic results were pages that name their own publisher as the best choice. Pro Marketer’s list of the top five AI SEO agencies in Canada puts Pro Marketer first. Run Marketing’s list of the top five AI SEO companies in Canada puts Run Marketing first. Julian Goldie’s site opens by naming Julian Goldie the best AI SEO consultant of the year, and to its credit it says plainly that the site is his. A ranked list written by one of the entrants is an advertisement with a numbered layout. Read it as one.

The more useful part was the question buyers ask: what makes an AI SEO consultant different from a regular SEO expert. That is the buyer’s real question, and it deserves a straight answer rather than a list.

What Google says the work requires

Google publishes a page called “AI features and your website” for site owners. Three sentences in it do most of the work of this post.

• “There are no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary.”
• “You don’t need to create new machine readable files, AI text files, or markup to appear in these features. There’s also no special schema.org structured data that you need to add.”
• Sites that appear in AI Overviews and AI Mode “are included in the overall search traffic in Search Console”, reported inside the ordinary Performance report under the Web search type.

To be shown as a supporting link, the same guide says, a page has to be indexed and eligible to appear in Google Search with a snippet. That is the ordinary bar. So for Google, the foundation of AI SEO is plain SEO done properly: a site that can be crawled, content in text rather than locked in images, structured data that matches the visible page, a Business Profile that is up to date.

That guide speaks for Google only. ChatGPT and the other assistants publish no equivalent promise, and I would not trust anyone who claims to know their selection rules in detail.

What genuinely changes

Something does change, and it is worth paying for if the consultant can show it.

When I measured eleven Canadian hiring searches in September, the industry searches carried an AI-written answer on nearly every run while the city searches carried a map pack instead. The full table is in what AI answers did to hiring searches. The consequence for a business is that some of its searches are now answered before anybody clicks, and on those searches ranking under the answer is worth less than it was.

That moves part of the job from “rank this page” to “be the kind of source the answer is built from”. In practice that means three things:

1. Pages that state something specific. A number with a source, a dated measurement, a method somebody else could repeat. A generic service page gives an answer nothing to quote.
2. Being described accurately elsewhere. An assistant asked for a recommendation draws on what other sites say about you. For a software product that is reviews and comparison coverage, which is why the SaaS SEO consultant page treats the AI answer as the main problem in that industry.
3. Measuring the answers repeatedly. An AI answer can appear on one run of a query and be gone on the next. One check tells you almost nothing.

None of that replaces the fundamentals. On a search with a city in it, the map pack still sits above everything, and that is profile and proximity work, the same job a local SEO consultant has always done.

Four claims to test before you hire one

These are the four claims I would test hardest, because each one can be checked on the first call.

1. “We will set up your llms.txt and AI schema.” Ask what that changes in Google. Google’s guide says no AI text file and no special structured data is needed for its AI features. A consultant may still have a reason to add a file for another assistant, and a good one will say it is a small experiment rather than the deliverable. If the file is the centre of the proposal, the proposal is thin.

2. “You will get a monthly report of your AI Overview traffic.” Ask where the number comes from. Search Console counts those clicks inside the ordinary Web search total, not as a separate line. A figure labelled as AI Overview traffic was estimated by some other tool, and you should be told which tool and how it estimates.

3. “Here is a screenshot of a client being named.” Ask how many times the question was run, and on which dates. My own measurement found queries where the answer changed between runs on the same afternoon. A single screenshot proves the business was named once. Ask for the same question run several times across weeks, with the count of runs that named the client.

4. “We will get you recommended by ChatGPT.” Ask what happens if it does not. Nobody controls which business an assistant names, in the same way nobody controls a Google ranking. A consultant can improve the odds and measure the result. A promise of placement is a promise nobody can keep, and I will not make it either.

A good answer to all four sounds similar: here is what I will change, here is how I will measure whether you are named, here is how often, and here is the month we decide whether it worked.

Short answers to the questions people ask alongside it

What is AI SEO called now? It goes by several names: GEO for generative engine optimisation, AEO for answer engine optimisation, and LLM optimisation. The first organic result on the Toronto version of this search uses GEO throughout. The names describe the same work, and none of them is a standard.

Can you do SEO with AI? Partly. Language models are useful for sorting long keyword lists, summarising a crawl and drafting first versions. They cannot run the measurement, and they cannot decide which of forty faults is costing you calls. That judgment is the part you hire a person for.

How much does AI SEO cost? In my practice it is not priced separately, because it is not a separate job. The work and the fees are set out on my SEO consulting services page, starting with a $499 audit, and the wider market is in what an SEO consultant costs in Canada. If someone quotes AI SEO as an extra line on top of an SEO retainer, ask what is in that line that is not in the retainer.

Is SEO still worth it in 2026? It depends on your demand and your margins, not on the calendar. The arithmetic is in whether an SEO consultant is worth it, and it works the same way whether the click comes from a classic result or a link inside an AI answer.

If you are hiring in Toronto

The Canadian search above was national. A Toronto business looking for help usually searches with the city in the query, and that page looks different: no AI answer on any of my September runs, three map pack slots, and two individual consultants ahead of every agency. If you are looking for an SEO consultant in Toronto, that page records who held the first page when I measured it and what the work costs if you hire me remotely. The AI answer work described here is part of it, not an extra.

Method

One query, “ai seo consultant”, pulled from the DataForSEO live SERP endpoint on 6 October 2026: desktop, English, location Canada, top ten results, a single pass. The Toronto details come from a second pull that day with Google’s location set to Toronto. I read every result page myself to decide who publishes it and what it sells. The Google quotations are from its Search Central page “AI features and your website”, read that day. A single pass over a page carrying an AI Overview is a snapshot, so treat it as one. Different results on your screen are worth an email: lasse@lassepettersen.com.