AI CitationTracker
How-to

How to track AI citations manually

You can learn more about your AI visibility in an hour with a spreadsheet than with most dashboards — if you track the right things consistently. Here's the process we use.

Step 1 — Build a fixed prompt set

Write 10–20 questions your buyers actually ask, phrased the way they'd ask an assistant. Include comparison questions ("best X for Y"), problem questions ("how do I..."), and brand questions ("what do people say about [category]"). Keep the wording fixed — you're measuring change over time, so consistency matters more than cleverness. Need ideas? The free audit tool will generate a starter set for your brand.

Step 2 — Pick your engines

Use the ones your audience uses: typically ChatGPT, Perplexity, Gemini and Google AI Overviews. Add Claude or Copilot if relevant. Check the same prompts in each engine on the same day.

Step 3 — Log what happens

For each prompt and engine, record:

Run each prompt two or three times and note variation — AI answers are not deterministic, and a single run is a weak signal.

Step 4 — Read the trend

After three or four weekly runs, patterns emerge: prompts where you're consistently absent, sources that keep appearing, competitors who show up often. Those are your priorities. For how to interpret each pattern, see step 4 of the audit checklist.

When to switch to a tool

Manual tracking gets painful with many prompts, daily cadence, several engines or client reporting. That's the point to compare software. Our tool comparison groups options by agency, in-house and content-team use cases. Understand the underlying metrics first in the citation explainer.