The Real AEO Playbook Lives in Content Strategy
by Lisa Peyton and her team of AI superheroes
Everyone has a tool, a tactic, or a framework right now for getting your content cited by AI. AICarma. DeepRanking. FoundInAI. llms.txt files. Schema markup. FAQ stuffing. The conversation about AEO and GEO has become a hype industrial complex, with most of the loudest voices selling the same SEO-imported tactics in new packaging.
I’ve been fielding questions about showing up in AI chat results for years now. These days, nobody’s asking if AI search matters. The real question is whether any of these so-called solutions actually move the needle.
So I decided to find out.
AI search isn’t search. It’s synthesis.
When a buyer, a client, or your VP types a question into Claude, ChatGPT, Perplexity, or Gemini, they don’t get ten blue links to pick from anymore. They get one answer, stitched together from a handful of sources the model decided to cite.
The old question was: how do I rank? The new question is whether you get cited at all.
That sounds like a small reframe. It isn’t. It changes the job.
Most of what’s being sold as “AEO” is the wrong answer.
A lot of what’s being published under the AEO and GEO banners is coming from SEO people who imported their tactical, manipulative mindset into the AI search era. The marketers I work with are exhausted by it. And the research actually shows most of these tactics don’t work.
An analysis of 300,000 domains found that having an llms.txt file has zero measurable effect on AI citation. Top 1,000 websites: zero adoption. The largest empirical study of AI citation predictors found that promotional tone correlates negatively with citation at 26 percent. The “optimize your way in” playbook is being sold on assumptions that the research is actively contradicting.
There’s a different version of this conversation that’s barely being held. The real question is whether your content can be cited, trusted, and surfaced when AI agents are doing the mid-funnel research your buyer used to do. That’s a strategy question, not an optimization question. It’s about authority, perspective, citability, and editorial judgment.
That’s our wheelhouse, content strategists. It’s not SEO 2.0.
I tested my own brand. Here’s what happened.
I scored three of my pieces against a five-characteristic Citation-Readiness Rubric I built. My Forbes article on AI shifts for marketing teams scored 10/10. My blog piece on content marketers moving upstream scored 9/10. My Maven course page scored 7/10. Then I ran two queries across the four major AI models, one about AI strategy for content marketers and one about which AI course to take.
My name appeared in zero of eight responses.
My course appeared in two (Gemini and ChatGPT named it as a top recommendation for the course query). The rest of the citation pool across all eight responses was Wikipedia, Reddit, vendor blogs, big SaaS company content, and one Harvard Extension School listing.
The most counterintuitive finding: my Forbes piece, which scored 10/10 on the rubric and lives on one of the most recognized business domains in the world, didn’t surface anywhere. Why? Forbes blocks AI crawlers. The piece is gated from Anthropic’s web fetcher, was central to Perplexity’s recent scraping lawsuit, and isn’t reliably accessible by the models that should be citing it. A 10/10 article on a top-tier domain is structurally invisible to the very models doing the synthesis. Domain authority doesn’t help if the model can’t actually reach the content.
The rubric works. The bigger problem is the pool.
Citation behavior is gated by source pool inclusion, which depends on factors above the piece level. Domain authority. Topical density. AI access posture. Earned media distribution. The research backs all of this. Brand web mentions correlate three times stronger with AI visibility than backlinks do (Ahrefs, 75,000-brand study). Eighty-two percent of AI citations come from earned media (Muck Rack, one million citations). Same content placed on a recognized outlet gets 325 percent more citations than the same content on a personal site (Stacker/Scrunch controlled study).
The strategic move isn’t “be more citation-worthy at the piece level.” It’s understand which pool your category pulls from, then decide what to do about it.
Which pool does your category pull from?
Almost nobody is asking this question, and it’s the most important strategic question in AEO right now.
Every category has a distinct source pool. For “best AI marketing courses,” the pool is course aggregators, big LMS platforms (LinkedIn Learning, Coursera, HubSpot Academy), and brand-name educational institutions (Harvard Extension, Northwestern Medill, eCornell). For “how should marketers adapt to AI,” the pool is Gartner, CMI, Harvard, arXiv, Adobe newsroom, and major tech vendor blogs. Different category, different pool. Different pool, different strategy.
You find your pool by running test queries in the four major AI models and capturing what each one cites. Same queries your buyer would actually ask. Both substantive (early-research questions) and commercial (ready-to-buy questions). Different phrasings to triangulate. Then you look at the patterns. Which domains keep appearing? Which platforms dominate? Which kinds of voices get represented and which don’t?
Once you know your pool, you have three strategic moves:
- Try to enter the pool through earned media placement, pitching the publications already in it.
- Build brand mentions across distributed channels (podcasts, peer posts, newsletters, communities), since AI synthesis reads prevalence, not just authority.
- Become a source by publishing original research, comparative tests, and distinctive arguments your peers will reference.
Most content marketers are choosing option 1 by default without checking their pool first. That’s why most AEO advice fails. The pool you should be optimizing for might not even include the kind of work you produce.
I’m walking through the full method live on Thursday, June 18th at 9am PT in a free Lightning Lesson on Maven: AI for Marketing: Make Your Content Citation-Worthy.
What we’ll cover on June 18th
You’ll leave with three things:
- A meta-prompting method for generating test queries across Claude, ChatGPT, Perplexity, and Gemini, so you can audit where your brand shows up in AI answers (and where it doesn’t).
- The Citation-Readiness Rubric, applied to real examples live in the session, plus the proposed sixth characteristic the test surfaced.
- The strategic framework for the layer above the rubric: how to think about source pool inclusion, earned media for AI citation, and the difference between being cited and shaping the synthesis.
If you’ve been quietly wondering whether any of the AEO tools and tactics being sold actually work, this is the room. Bring your own brand. Run the test alongside me.
