Every time someone walks me through the new AEO playbook, I recognise it. You get mentioned on other sites, answer the question properly, cover the follow-ups, show up on Reddit and YouTube, and make it obvious who you are.
That’s an SEO plan, and it was one in 2014 too. Good SEOs were selling it long before anyone called it AEO or GEO.
I don’t mind the new name much. What bugs me is the pitch that it’s a new discipline, with a new budget, that your SEO team can’t do. Most of it is the work they’ve been asking to do for years. Plenty of very experienced SEOs, Ian Howells among them, say the same.
Google says the same. Its guide to its AI features says optimizing for them “is optimizing for the search experience, and thus still SEO.” AI Overviews and AI Mode pull their sources from Google’s search index, so that’s simply how they work.
AEO tactics SEOs were already using
Go through any AEO checklist and most items have an obvious older version.
Answer the question first. Dr. Pete wrote this up for featured snippets back in 2016. Rank on page one, then lead with “a quick summary at the top, followed by a deeper dive.” The main new advice is to chop pages into tiny chunks for LLMs, and Google’s own AI guide says you don’t need to. Clear sections that each answer one question, with the answer up top, still help any engine pull the right snippet.
Cover the questions people ask next. Query fan-out means the AI runs a handful of extra searches for the follow-up questions, so your page has to answer more than the headline query. We used to call that covering search intent, and AJ Kohn put it well in 2015: “Target the keyword, optimize the intent.”
Get people talking about you. Rand Fishkin did a Whiteboard Friday in 2012 about Google noticing brands that get mentioned next to a topic, even without a link. Now Ahrefs finds brand mentions line up with AI Overview visibility far more than backlinks do.
Be where your customers already talk. The old BlueHatSEO e-commerce checklist from 2010 has most of it. Join the forums in your niche and answer product questions every week, put your products in Google’s shopping feed, film reviews and post them on the video sites. Swap in Reddit, Merchant Center and YouTube and you’ve got a 2026 AEO retainer.
Get onto the lists that get cited. The “citation gap” report every AEO tool sells means checking which pages the AI cites and getting your brand onto them. That’s link building from your competitors’ backlinks, which SEOs have always done.
Make it obvious who you are. Clear About page, consistent name and description everywhere, structured data, a fair Wikipedia entry if you’ve earned one. AJ Kohn wrote the checklist in 2014 and called it Knowledge Graph Optimization. His rule still holds: “Write clearly so that both humans and search engines know what the hell you’re talking about.”
And the basic one: let the crawlers in. OpenAI, Anthropic and Perplexity don’t publish ranking guides. What they do publish is “don’t block our search bot”. That’s robots.txt, which is about as old as SEO gets. As for the newer llms.txt file, Google’s AI guide says you can ignore it for Google Search.
AEO spam is old SEO spam on new platforms
The shady AEO tactics are the shady SEO tactics with new platform names.
Fake Reddit accounts are the 2026 version of spamming Yahoo Answers. A page for every fan-out query is the million-page database site. Planting brand facts on Wikipedia is, well, planting links on Wikipedia.
They died in Google for the same reason they’ll die here. Google’s AI guide calls mass-produced fan-out pages scaled content abuse. The tactics that lasted were the ones that left something real behind, like a question actually answered, a review someone actually filmed, a product feed that’s actually accurate.
Self-ranking “best X” lists are the awkward one, because they work right now. When ChatGPT cites a software company’s own site for a buying question, Ahrefs found it’s a “best” list on the company’s blog about a third of the time. I don’t know how long that lasts. Like every SEO play, there’s a credible version and a spammy one. The credible list tests each option, says who it suits and ranks a competitor first when it’s better. The spammy list puts the publisher at number one every time, which is voting for yourself on a top sites list, and that’s the version I’d expect to stop working first.
Where AEO and SEO actually differ
There’s no rank to track. Ask an AI the same question twice and you’ll often get a different list of brands. SparkToro ran this at scale and found the same list came back less than once in a hundred tries. So anyone selling you an “AI ranking position” is selling something that doesn’t exist. You measure by asking the same questions many times and counting how often you show up.
People don’t type keywords. Prompts are long and personal. The engine turns them into its own searches that you can’t see, so your keyword research is now a guess at what the machine will search for.
Other people’s pages matter more. AI answers often cite reviews and “best” lists, which Ahrefs found make up almost half the pages ChatGPT cites for buying questions. More of the work happens off your site.
Each engine is its own thing. ChatGPT, Google’s AI Mode, Perplexity and Claude pull from different places. Ranking well in Google doesn’t mean ChatGPT will mention you, so you check each one.
None of that adds up to a new set of tactics, though. It’s a new way of measuring, plus more weight on stuff SEOs already did.
Why AI search matters while it sends less traffic than search
Across 118,671 sites in Ahrefs’ October 2026 data, AI assistants sent 0.46% of visits. Search engines sent 27.55%, about 60 times as many.
- Search engines: 27.55 % of visits
- AI assistants: 0.46 % of visits
That gap is real, but I don’t read it as a reason to ignore AI. Ahrefs counts clicks, and when an agent reads your page and answers without sending anyone, no visit is recorded. Referral share undercounts how much AI already uses your content.
Human visits are also falling while AI use grows. Visits to the top million sites are down about 12% since 2022 while people online grew about 24%, so visits per person fell about 29% (SparkToro). Rand Fishkin is clear that they can’t say for certain why: apps, AI tools and search keeping people on their own pages, and the end of Covid lockdowns, could all play a part. Similarweb’s browser panel can’t see an agent reading a page either.
My bet is that agents are a big part of it, and that their share grows fast. Within the next two to three years I expect AI agents to do most of the informational searching SEO has always targeted, with AI use growing 10 to 100 times. Entertainment, like YouTube, will stay more human. That’s a prediction, and the data doesn’t prove it yet.
So AI search isn’t a distraction, and it isn’t a reason to move budget out of SEO either. The SEO work is what gets you into the agents’ answers, so it’s the same work with higher stakes.
What I’d do instead of buying AEO
Here’s where I’d put the time and money:
- Check robots.txt and your CDN aren’t blocking Googlebot, Bingbot or the AI search bots.
- Find the questions your buyers actually ask, and make sure a page answers each one in its first sentence and covers the follow-ups.
- Get talked about in the places they hang out, like the review sites, the lists, the subreddits and the YouTube reviews, using real accounts and a product people actually like.
- Make your About page and profiles say clearly who you are and what you do.
- Once a month, ask the AI engines your 10 most important questions a few times each and note how often you come up.
The first four are what SEOs have been asking for budget to do for a decade. If you’ve been doing them all along, you’re already doing AEO. Tell whoever’s pitching you exactly that.