The AEO.ca Method, in four numbered parts
Nobody has AI search figured out, including us, so the Method is a baseline and evergreen work against it: pages that argue your case for a specific person, an ecosystem that corroborates it everywhere the agents go, and one clear statement of what you do and for whom, measured on an instrument we built and checked against a second one.
Written by Michael Landry, Founder and Head of AEO.
Where AEO and SEO meet
AEO and SEO overlap, probably by about 70%, and every SEO tactic still applies. SEO is getting Google and the other search engines to understand your website as the answer to the query someone typed in, and it has mostly been done with keywords. AEO works with phrases instead, because every phrase can be asked a dozen ways and expanded, which is what the engines call fan-out. We plan around personas and what matters to them, not around a keyword list.
The bigger difference is the job of the content. With SEO you want to be the authority on a topic, which is to say you want to be Wikipedia. With AEO you want to be the best answer for that specific person, and prove it, which is to say your content has to be framed as an argument: why you are the best choice, what your value proposition is, what you guarantee, how you stack up against the alternatives. Every person now has an agent that goes out, does the research and fetches answers for them, and that agent wants to come back with the best answer for its user.
1.0 Be the answer
SEO has over 200 signals, and about five percent of them get you ninety-five percent of the result: the quality of the written content, page speed, internal linking and inbound links. That foundation comes first on every plan. On top of it we build content planned by persona and phrase, framed as the case for why you are the best choice, with your USPs, guarantees and comparisons, opening with the answer and marked up with structured data that says the same thing the page says. Everything we publish reads at least 80% human by Grammarly’s measure, because an agent gets nothing from content it could have written itself.
2.0 Be recommended
Good AEO means providing the proof and the answers not only on your website but everywhere that matters: social, forums, review sites, award sites, natural PR and earned media, video and imagery. The useful thing about an AI query is that when the agent goes out to put the pieces together for its user, it tells you where it went. If we can see where it went, we can see your gaps. For each client we map who the customers are, what matters to them, where the agents go when those things get asked on Perplexity or Claude or ChatGPT, and whether you are there. Then we plant your flag in as many of those places as makes sense, including applying for awards on your behalf, because an award is social proof an agent can read.
3.0 Be understood
One clear statement of what you do, for whom and where, carried everywhere an agent reads: your site and its structured data, Business Profiles, LinkedIn, directories, press boilerplates, creator briefs, video descriptions. We call it your Answer Profile, and it is the first deliverable in every engagement.
4.0 Measurement
Visibility is hard to see at scale, and this is a new industry. What we do is take a baseline and work against the relative movement. Our instrument of record is the tool we built, because we know exactly how it was built and what goes into every report. We benchmark against Semrush as a second source, because two sources keep each other honest: if both show progress by their own method, you can be reasonably assured; if one shows growth and the other shows loss, we investigate rather than pick the flattering one. Reporting is monthly, against the twelve-month plan we agreed on in month one.
Month one
Month one on every plan is setup, and the bigger the plan, the more setup there is: research, persona identification, technical analysis and recommendations, phrase research for those personas, opportunity finding, and a plan for the next twelve months that everyone has agreed to. We do not call it an audit, and there are no setup fees.
What Google says, and how we stay inside it
Google’s own guidance is that there are no special requirements for AI Overviews or AI Mode beyond the fundamentals, that structured data has to match the visible text, and that no special file or markup is needed. We build to that. Files like llms.txt and open access for the other engines’ crawlers are hygiene for ChatGPT, Perplexity and Claude, not a Google lever, and we say so.
The part that will keep changing
How people use AI, how it operates and how it delivers answers will keep changing. All anyone can do is take a baseline and work against it, and the only things with lasting value are tactics that serve your customers in an evergreen way. More than anything, you want to be a good business, because there is a limit to how much can be faked, and AEO is going to reward real businesses with real happy customers.