AI visibility starts with the questions your clients' customers actually type into ChatGPT, Gemini and Perplexity. We agree a list per client with you, covering the service, the location and the comparison questions buyers ask. The agent then checks how each assistant answers and records whether the client is named, described correctly and linked.
We keep the list short at first, usually the questions that decide a purchase, and widen it once the baseline is clear. A long list of vague questions only produces noise. Run the same questions on each assistant, because they can answer differently. The agent records each one separately, so you can see where a client is strong and where it is invisible.
When a client is missing, the agent looks at which sources the assistants appear to rely on, such as directories, review sites, publications and the client's own pages. It then proposes content and corrections to earn a place there. Your team decides what gets written, pitched or updated, and the agent tracks what changes over the following weeks.
AI answers change from one day to the next, and nobody controls what an assistant says. The agent shows trends, not promises. Treat early results as a baseline, then compare month on month. It is a poor fit for a client with no website or content to build from. Include the findings in your regular reports so clients see the new service line.
ChatGPT, Gemini, Perplexity, Claude and Copilot, with more added as they matter.