How to Measure Your AI Visibility (DIY Guide)

Measure your AI visibility by running a fixed set of your customers' real questions across the major engines on a schedule and logging who gets cited. You can do it manually for free or automate it with a tracker — the method is the same, and the metric that matters is citation share, not rankings.

Step 1 — Build your prompt set

Start with the real, conversational questions your customers ask AI — not keywords. Pull them from sales and support conversations, Reddit and community threads, and by prompting the engines yourself. Aim for 10–30 high-intent questions, weighted toward the ones closest to a buying decision. Keep this list fixed so your readings stay comparable; add to it deliberately, don't churn it.

Step 2 — Run them across the engines

Run each prompt across the engines your audience actually uses — typically ChatGPT, Perplexity, Google AI Overviews, Gemini, and Microsoft Copilot. Because generation is non-deterministic, run each prompt a few times (or rely on a tool that samples daily) rather than trusting a single response. Do it on a fixed cadence — weekly is a reasonable manual baseline; daily is better and is where tools earn their cost.

Step 3 — Log mentions, citations, and competitors

For each prompt and engine, record three things, because they're different signals:

1. 1

   Mention

   Did the answer name your brand at all? This is presence in the response.

2. 2

   Citation

   Was your site linked or used as a source — even if your brand wasn't named? This is the prize: the click that survives the AI answer.

3. 3

   Competitors & sources

   Which other brands were named, and which domains (Reddit, review sites, publications) the engine cited. Gaps where competitors appear and you don't are your roadmap.

A simple spreadsheet — prompts down the side, engines across the top, a cell for mention/citation each run — is enough to start.

Step 4 — Track citation share over time

Turn your log into a trend. Your headline metric is citation share: across your prompt set, how often are you cited versus competitors, per engine? Watch the direction over weeks, not the absolute number on any day — because AI citations are volatile. Semrush found 40–60% of LLM-cited sources change month to month, and Google keeps a given URL in an AI Overview only about 3.87 days on average. A single reading is noise; the trend is signal. This is the Adaptability pillar in practice. For competitor tracking specifically, see how to know if AI is citing your competitors.

Step 5 — Add AI referral and brand signals

Round out the picture with two downstream metrics. In your analytics, watch referral traffic from AI domains (chatgpt.com, perplexity.ai, gemini.google.com) — it's small but high-intent and growing. And watch branded-search lift: when engines name you more, more people later search your brand directly. Together with citation share, these tell you whether your AEO work is compounding.

When to use a tool instead

Automate when manual tracking starts hiding the picture — when you need daily sampling, more than a handful of prompts, several engines, or competitor and source analysis at scale. Free tiers cover a baseline; paid trackers handle the volume. Compare your options in the best AI visibility tools and the best free AEO tools.

The bottom line

You don't need a budget to measure AI visibility — a fixed prompt set, the major engines, and a spreadsheet get you a real baseline, and free trackers automate it. Track citation share per engine over time, act on the gaps, and re-measure. Start from what is AEO for the why behind the metric.