How to report AI visibility to clients (with a template)
A practical guide for agencies: what goes in a monthly AI visibility report, the metrics clients understand, a template, a checklist and how to explain changes.
Short answer: A good AI visibility report answers four questions on one page: where does the client stand, who is beating them, what changed this month, and what happens next. Lead with one score and its trend, show the competitors AI names instead, list the buyer questions won and lost, tie it to traffic and leads where you can, and finish with the work done and the plan. Use the same structure every month, and explain up front that AI answers vary, so the trend matters more than any single result.
Clients now ask their agencies a new question: “Does ChatGPT recommend us?” Answering it once is easy. Reporting on it every month, in a way a busy client understands and trusts, is where agencies struggle. This guide gives you a report structure you can reuse, the metrics worth showing, a checklist, the words to explain fluctuation, and straight answers to the questions agencies ask about tooling and pricing.
What should an AI visibility report include?
An AI visibility report should include a headline score with its trend, a breakdown for each AI engine, a competitor leaderboard, the buyer questions won and lost, what AI says about the brand, the business impact, the work completed and the plan for next month. Nine short sections are enough, and most fit on a single page.
| Section | What it shows the client | Where the data comes from |
|---|---|---|
| 1. Headline score | One number for this month, against last month and against the starting point | Your AI visibility platform |
| 2. Engine breakdown | How the brand performs in each engine, so weak spots are obvious | Your AI visibility platform |
| 3. Competitor leaderboard | Who AI names instead, the client’s rank, and who moved | Your AI visibility platform |
| 4. Wins and losses | Three to five buyer questions gained and lost, with the answer text | Tracked prompts and saved answers |
| 5. What AI says | The tone of the descriptions, and any facts AI gets wrong | Saved answers |
| 6. Sources | The sites the engines name or link in their answers | Saved answers |
| 7. Business impact | Visits from AI assistants, branded search and enquiries that mention AI | Analytics, Search Console, the client’s CRM |
| 8. Work completed | What you shipped this month, with dates | Your task list |
| 9. Next month | The three changes most likely to lift the score | The platform’s fix list and your judgement |
Keep the order fixed. A client who reads the same layout every month learns where to look, and stops asking you to walk them through it.
Which metrics do clients actually understand?
Clients understand a score out of 100, a ranking against named competitors, and a short list of questions won and lost. They rarely need engine level detail or raw answer logs, so keep those in an appendix and lead with the numbers a managing director can repeat in a meeting.
| Metric | How to explain it | Why the client cares |
|---|---|---|
| AI visibility score | “How often and how well AI recommends you, out of 100” | One number to watch |
| Mention rate | “AI named you in 14 of the 25 questions your buyers ask” | Concrete and easy to picture |
| Position | “When you are named, you are usually second on the list” | First place gets most of the attention |
| Share of voice | “You get a quarter of the mentions in your category” | Shows market standing |
| Competitor rank | “You are third, behind two named rivals” | Makes the gap personal and specific |
| Sentiment | “AI describes you as reliable but expensive” | Shows what to correct |
| AI referral visits | “83 people clicked through from AI assistants” | Links visibility to traffic |
If your platform gives a single score, use it as the headline. ONEGO’s is the ONE Score, and our page on the AI visibility score explains how to read one. Avoid comparing scores from two different tools in the same report, because each vendor calculates its score differently.
A reporting checklist clients can follow
Use this checklist before each report goes out. It keeps the report honest, consistent and easy for a client to read.
- Same prompts as last month. Only change the tracked questions deliberately, and say so when you do.
- Same layout as last month. Do not move sections or swap charts.
- Lead with the trend, not the single latest reading.
- Name the competitors. “Two rivals” means nothing; real names do.
- Show one real answer. Paste a short AI answer that names the client, or one that leaves them out.
- Flag wrong facts. If AI states an old price, address or service, list it with the fix.
- Separate what you did from what happened. List the work, then the change, and be careful about claiming cause.
- Include one business number, even if it is small: AI referral visits, branded searches or enquiries.
- State the limits. One line saying AI answers vary and results are indicative.
- End with three actions and who owns each.
How do you explain week to week changes?
Explain that AI answers are not fixed. The same question can produce a different answer minutes later, and results differ between users and locations, so a small move in either direction is normal. What matters is the direction over several weeks, measured with the same questions each time.
Say this at the start of the engagement, not the first time a number drops. A short note you can reuse in every report:
AI engines generate a fresh answer each time, so results vary from run to run. We track the same buyer questions every week and report the trend. A change of a few points in one week is normal; a change that holds for a month is real.
When something does move sharply, check three things before you report it: whether one engine or all of them changed, whether a competitor gained at the same time, and what you or the client shipped in the weeks before. Our AI visibility tracker page covers how to read these patterns.
How often should you report?
Report to clients monthly, review the data yourself weekly, and hold a deeper review each quarter. Monthly is frequent enough to show progress and slow enough for the trend to mean something.
- Weekly, internal: check for sharp drops, wrong facts and competitor moves. Act on them; do not send a report.
- Monthly, to the client: the one page report above.
- Quarterly: revisit the tracked questions, the competitor set and the plan, and agree priorities for the next three months.
How do you tie AI visibility to business results?
Tie AI visibility to business results with three numbers: visits referred by AI assistants, growth in branded search, and enquiries where the buyer says AI sent them. None is complete on its own, so show them together and be clear that AI visibility is a leading indicator, not a guarantee of sales.
- AI referral visits. In Google Analytics 4, filter sessions by source with a pattern that matches the main AI assistants. The exact pattern is in our guide to tracking brand mentions in ChatGPT. Treat the number as a floor, because some AI traffic arrives with no referrer.
- Branded search. Buyers who see a brand in an AI answer often search for it by name later. Watch branded impressions and clicks in Search Console. Google says traffic from its AI features is counted in overall Search Console data rather than reported separately (Google Search Central), so you will not see AI Overviews as their own line.
- Enquiries. Ask the client to add “How did you hear about us?” to forms and sales calls, with an AI assistant option.
Never promise a ranking, a number of leads or a revenue figure from AI visibility work. Report what you measured and what you did.
Can you manage multiple clients from one dashboard?
Yes, if the platform’s plan covers several brands. Look for one account that holds every client as a separate brand, with its own questions, competitors and history, plus enough seats for your team and reports that can carry your agency’s branding.
On ONEGO, the Growth plan covers 5 brands and 3 seats with 50 tracked prompts, and the Agency plan covers 10 brands and 10 seats with 100 tracked prompts, white label and API. Every plan refreshes 4 times a month across all five engines. See ONEGO for agencies and the checklist for choosing an AI visibility tool for agencies.
Should you pay for each client brand, or put all clients on one plan?
For most agencies, one plan that covers several brands is simpler and cheaper than paying for each client separately. Paying brand by brand makes sense only when you have one or two clients, or when each client needs a very large set of tracked questions.
AI visibility platforms charge in different ways, and the model changes your margin more than the headline price does:
| Pricing model | How it works | What to check |
|---|---|---|
| By brand | A fee for each client brand you add | Cost when you reach 5, 10 and 20 clients |
| By prompt | A pool of tracked questions shared across brands | Whether the pool is enough for every client |
| By seat | A fee for each team member | Whether clients need their own logins |
| By engine | Extra engines cost more | Which engines are in the plan you would buy |
| By credit | Each check uses credits | How many checks a typical client needs each month |
ONEGO prices by plan rather than by client: the Agency plan is 299 US dollars a month, or 239 dollars a month billed yearly, for up to 10 brands sharing 100 tracked prompts. Full details are on the pricing page. Whichever platform you choose, work out the cost for the number of clients you expect in a year, not the number you have today.
How do you use a free AI audit to win a new client?
Run a free AI visibility audit on the prospect’s domain before the pitch, and open the meeting with what you found: their score, the competitors AI names instead, and two or three questions they are missing from. A specific gap with a named rival is far more persuasive than a general talk about AI search.
A simple sequence that works:
- Before the meeting, run the audit and note the score, the top competitor and one answer that leaves the prospect out.
- In the meeting, show that one answer. Let them read what their buyers see.
- Propose a first 90 days: fix crawler access, clarify the key pages, earn mentions on the sources AI cites, and report monthly.
- When they sign, add them as a brand on your plan so tracking starts from a clear baseline.
ONEGO’s free audit shows a ONE Score and report with no card needed, so it is a low friction way to start that conversation. Our playbook for agencies covers the service line in more depth, and the free AI visibility checker explains what the audit measures.
Common reporting mistakes
A few habits make AI visibility reports harder to trust.
- Reporting a single good answer as a win. One answer is one sample. Report the pattern.
- Changing the questions every month. The trend becomes meaningless.
- Hiding a drop. Clients forgive a dip that is explained. They do not forgive a surprise.
- Claiming credit for every rise. Engines change on their own. Say what you did and what moved, and let the pattern build over months.
- Sending a dashboard login and calling it a report. Most clients want your reading of the numbers, not the numbers.