Picture a plumber in Tucson. Great reviews, years in business, a clean website. She's not showing up when someone asks ChatGPT or Gemini to recommend a local plumber. A competitor with half her experience is getting named instead — repeatedly.
That's the kind of gap that started this whole project.
When business owners first started noticing they were invisible in AI-generated recommendations, the standard advice was vague: "be consistent online," "get more reviews," "have a good website." True, but not useful. Nobody could tell you which part of your presence was causing the problem, or what to fix first.
So we decided to figure it out ourselves.
We didn't want to assume. We started by running dozens of real audits on businesses across different industries and cities — plumbers, lawyers, dentists, HVAC contractors, restaurants, consultants. We asked AI assistants the same questions a real customer would ask. We recorded who got recommended and who didn't.
Then we looked at every measurable difference between the businesses that showed up and the ones that didn't.
Not theories. Not educated guesses. Actual patterns from actual audits.
What emerged was a set of categories — clusters of signals that AI assistants appear to weigh when they decide which local business to name. We kept testing. We kept finding new variables. We kept refining what was genuinely predictive versus what was just noise.
That process is what gave us the AI Ranking Factors framework — and eventually, an index of more than 150 specific, measurable factors.
The index isn't a random checklist. Every factor fits into one of several core categories, each representing a different way AI systems evaluate your credibility and relevance.
AI assistants pull heavily from structured data — your Google Business Profile, directory listings, social profiles. An incomplete profile isn't just a minor gap. It's a credibility signal that cuts against you. We measure more than a dozen specific fields and attributes across the most influential platforms.
Volume matters. Recency matters. But so does something most people overlook: the language inside your reviews. When customers describe specific services, locations, and outcomes, they're giving AI systems something to quote and corroborate. We measure review patterns across multiple dimensions, not just your star rating.
AI assistants are essentially pattern-matching engines. The more places your business name, location, category, and credentials appear in consistent, trustworthy contexts — the more confident the AI becomes in recommending you. We call this corroboration, and it's one of the highest-leverage categories in the entire index.
Your website might look great to a human and be nearly invisible to an AI. We analyze how cleanly your site's content can be parsed, extracted, and understood — structure, clarity, semantic organization. A beautifully designed site with poorly organized text can actually underperform a plain, well-structured page in AI outputs.
Licenses, certifications, associations, awards, press mentions — these are trust anchors. We measure not just whether they exist, but whether they're surfaced in places where AI systems are likely to find and weigh them.
To make this concrete, here's a sample of the kinds of factors that show up across these categories:
None of these are secrets. But most businesses have never looked at their presence through this lens — and the gaps are usually significant.
The thing that frustrates me most about the conversation around AI visibility is how much of it stays abstract. "Build trust." "Be authoritative." "Have a strong presence." These aren't wrong, exactly — but they're not actionable.
The reason we built an index with 150+ specific factors is precisely to escape that abstraction. When you can measure something, you can prioritize it. When you can prioritize it, you can actually improve it. And when you improve it, you can measure the change.
That's the loop that matters: scan, identify, fix, rescan.
Some factors move quickly — a profile update, a structured page rewrite, a targeted ask for a detailed review. Others take time to compound. The index helps you see which is which, and where your time is actually worth spending.
AI assistants are not static. The models change. The sources they draw from shift. New platforms gain weight; old ones fade. The index we use today isn't identical to the one we started with, and it won't be identical to the one we use a year from now.
That's not a weakness — it's the point. This isn't a one-time certification. It's an ongoing discipline.
The plumber in Tucson? She had strong reviews but almost no corroboration outside of two platforms, a half-filled business profile, and a website that was beautiful on mobile and nearly unreadable to a language model. Three targeted fixes later, she started showing up. Not overnight — but measurably, trackably, repeatedly.
That's what a real index is for.
— J. Brent Tuttle