Machines don’t rank you.
They understand you…or they don’t.
AEO isn’t a channel, it’s infrastructure. Entity graphs, relational markup, and off-site presence. That foundation determines whether AI systems cite your business or skip it entirely.
Measuring AEO: what it is, and isn’t.
Most SEOs (myself included) spent the first wave trying to measure AI search like a performance channel. Query volume, attribution models, conversion paths. The usual toolkit.
But the problem is, none of that works yet. There’s no query volume to optimize against. No real feedback loop. Attribution is a guess. A buyer makes up their mind inside ChatGPT, closes the tab, and buys somewhere else.
What actually moves the needle is boring. Coherent entity graphs. Structured data that tells a relational story, off-site brand mentions. That’s not keyword research. That’s not content gap analysis. That’s infrastructure.
SEO has always been about getting humans to click. AEO visibility may be the first time success depends on whether machines actually understand you.
This doesn’t mean measurement won’t catch up, it will. But right now, the teams that are winning are the ones building the foundation, not throwing all their eggs into a basket they don’t really know even exists yet.
AEO levers to pull. None of them are shortcuts.
There are new studies coming out every week that help us understand more about AEO/GEO. But the there are common pillars that we know matter when AI visibility is the goal.
LLMs build confidence from third-party signals like industry publications, niche forums, expert roundups, the sites they already trust as training sources. Your own website is confirmation, not the trigger. The relationships and editorial placements that built link equity and brand authority are the same ones that build LLM confidence. Be known, be credible, be where people talk about your category. The delivery mechanism changed, but the fundamentals didn’t.
Make entity relationships parseable. The internal link structure is the first signal of how your business, services, locations, and people relate to each other. If the architecture is incoherent, the entity graph is incoherent. No amount of off-site work compensates for a site that can’t explain its own structure.
Not page-by-page markup. Relational DNA. The structured data layer should tell a consistent story across an entire site, connecting the Organization to its Services, Locations, People, and content. When done right, it’s a machine-readable map of reality. When done wrong, it’s noise that actively confuses the systems trying to read it.
Entity-rich, not keyword-stuffed. Content that establishes what something is, who’s behind it, and why it matters. LLMs don’t count keywords. They map semantic meaning, and they’re looking for clear entity relationships to do it. The content needs to be support beams for those relationships.
Not a 12-step process.
A playbook.
Every engagement starts with the same question: what does your site look like to a machine reading it cold? Everything after that depends on the answer.
Map every entity your site declares: Organization, Person, Service, Location, etc, and how the properties connect. Find where the relational story breaks down. This is the diagnostic that shapes everything else.
Resolve contradictions, build the cross-page entity connections, establish the structured data foundation, make sure the ChatGPTs and Claudes of the world can even access your site. This area is where most sites have the biggest gap, and where the highest-leverage fixes typically live.
Restructure content so it reinforces the entity graph rather than floating independently. Every page should strengthen the overall story, not just target a keyword.
Earn the brand mentions, editorial placements, and third-party citations that make LLMs confident enough to cite you. This is where SEO and digital PR activities like link building overlap with AEO.
Results, credentials, and tools I’ve built.
Implemented sitewide AI-focused structured data to establish a relational entity graph for a large plastics manufacturer. The result was a 40% increase in visibility across AI-driven answer platforms.
Built and led Heroic Search, a remote-first SEO and link acquisition agency. Over a decade, developed a relationship database of site owners, journalists, and bloggers across every major vertical. Clients included Elementor, Benzinga, Freethink, Common Desk (acquired by WeWork), and Rebrandly.
Custom-built auditor using Claude Code