A client AI visibility report should open with a plain summary of how AI engines name the client, then show visibility and share of voice by engine and subcategory, the competitors that appear, the sources the engines cite, what changed since last time, and a short list of next steps. Keep counts beside every percentage and explain that answers vary, so trends matter more than any single result.
This guide lists the sections to include, how to present them and how to avoid the mistakes that cost agencies client trust.
What is an AI visibility report?
It is a regular report on how AI engines such as ChatGPT, Gemini, Google AI Overviews and Google AI Mode talk about a client's brand: whether they are named, how they compare with competitors and which sources shape the answers. It plays the role for AEO that a ranking report plays for SEO. For the metrics behind it, see AI visibility metrics.
Which sections should the report include?
| Section | What it answers |
|---|---|
| Summary | How are we doing, and what changed? |
| Visibility and share of voice | How often are we named, and how do we compare with rivals? |
| Subcategories | Where do we win and lose by topic? |
| Engines | Do ChatGPT, Gemini, AI Overviews and AI Mode differ? |
| Competitors | Who is named instead of us, and where? |
| Sources and citations | Which pages shape the answers, and where are the gaps? |
| Prompts doing well and poorly | Which questions gained or lost ground? |
| Concerns | What is wrong, outdated or risky in the answers? |
| Strategies and next steps | What should we do now? |
| Appendix | Every prompt tracked, for transparency |
How should you write the summary?
Lead with the answer, in three or four sentences:
- The headline number with its count, such as "named in 18 of 60 checks" (an example format)
- The direction since the last report
- The biggest win and the biggest gap
- The single most important next step
A busy client may read nothing else, so make it stand alone.
How do you present the numbers honestly?
- Show counts. A percentage without its denominator hides how much data stands behind it.
- Show trends. Compare against the baseline and the last period.
- Explain variation. OpenAI describes model output as non-deterministic (OpenAI), and research published by SparkToro and Gumshoe.ai in January 2026, not peer reviewed, found that repeated runs of the same prompt rarely returned identical brand lists (SparkToro). Say so in a sentence, and judge changes over weeks.
- Treat position carefully. Use average position as a direction, and lead with visibility.
How do you turn results into actions?
For each finding, give an action, an owner and a way to check it worked:
- Weak subcategory: add or improve pages that answer those questions plainly.
- Citation gap: correct or earn coverage on the third-party source. See AI citation gaps.
- Blocked crawler: fix robots.txt. See Can AI crawlers read your site?
- Wording sensitivity: cover both phrasings and track them. See Why AI answers change with wording.
How often should you report?
Daily checks feed the data, and most clients want a weekly digest or a monthly review. Weekly digests keep the client engaged between reviews, while a monthly report is the place for trends and strategy. Whatever you choose, keep the format identical so changes are easy to spot.
How does white-label reporting work in Verte?
Verte builds the report from your tracked data. Every report uses the same format: a summary, subcategories, competitors, citations, concerns and strategies, then an appendix of every prompt. The narrative is written from your data and the figures are computed, and you can refresh the summary before sending. With white-label on, the report carries your agency's name, logo and colours and no Verte branding. Clients can get a PDF or a share page, and weekly digests can go to client contacts.
See white-label reports and the Platform's report section. For the wider service around reporting, read AEO for agencies and How to pitch AEO to a prospect.
Start as an agency. See current plans.