"Are we showing up in AI search yet?" is the question every AEO client eventually asks, and it deserves a more precise answer than a shrug. Here's how to actually measure it.
Why This Is Harder Than Tracking a Google Ranking
A Google ranking is a fixed, checkable number: a page is #1, #4, or not in the top 100 for a given keyword, and that position is roughly stable from one search to the next. AI answers don't work that way. Ask the same question twice, even minutes apart, and an AI assistant can phrase the answer differently, name a different second or third business, or skip a citation it gave last time. This isn't a bug in the tracking, it's a property of how these systems generate responses. Measuring AI visibility means measuring a pattern across repeated samples, not checking a single fixed position.
The Core Metrics
AI Mention Share. The percentage of relevant prompts, run repeatedly across a defined panel of realistic customer questions, in which a business is named at all. This is the closest AEO equivalent to a keyword ranking, and it's the metric most worth tracking monthly.
Share of Voice vs. Named Competitors. Among the prompts where any business in the category gets named, what percentage of those mentions go to your business versus each specific competitor. This turns "are we visible" into "are we winning," which is the more useful question once baseline visibility exists.
Citation Sentiment. Not just whether a business is mentioned, but how. An AI assistant can name a business while also noting a caveat ("solid but pricier than competitors") or a limitation. Tracking sentiment alongside mention frequency catches problems that a raw mention count would miss entirely.
Prompt Coverage. How many distinct, realistic phrasings of a customer's question actually surface the business. A business that only appears for one exact phrasing has far less durable visibility than one that appears across a dozen variations of the same underlying question, because real customers don't all phrase things identically.
Source Attribution. When an AI system does cite the business, which specific page or piece of content is it pulling from. This tells you which content assets are actually doing the work, and which are getting ignored despite existing.
How to Actually Track These
Build a realistic prompt panel first. Not a handful of obvious questions, but 20 to 50 realistic variations of what actual customers ask, across different intents (comparison questions, "best of" questions, problem-specific questions, direct brand questions).
Run the panel across platforms, repeatedly. Test the same panel across ChatGPT, Perplexity, Gemini, and Google AI Overviews, and run each prompt more than once per platform, since answers vary session to session. A single pass through the panel produces a noisy snapshot, not a reliable measurement.
Log results consistently. Whether tracked manually in a spreadsheet or through a dedicated AI visibility monitoring tool, the discipline that matters is consistency: same panel, same platforms, same cadence, so month-over-month comparisons mean something.
Track monthly, review quarterly. AI visibility moves slower than a single content update, because it depends on authority and trust signals that compound over time. Monthly tracking catches trend direction; quarterly review is where it's worth reassessing strategy.
What "Good" Actually Looks Like
There's no universal benchmark number, because it depends heavily on competitive density in a given category and market. A business in a category with little AI-optimized competition can reach meaningful mention share quickly. A business competing against several well-established, AI-optimized competitors in a dense urban market should expect that same climb to take longer and require more sustained content and authority work. The trend line matters more than any single month's number: mention share and share of voice should be moving up, consistently, over a two-to-three-month rolling average.
Not sure where your numbers stand today? Get in touch and we'll help establish your baseline across all four major platforms, so you know exactly what you're improving from.