AI Search Visibility Reporting Tool: Client Reports for Agencies, Otterly, Peec.ai, Profound, Semrush
How we choose a reporting tool for AI visibility client work, what we expect each month's report to contain, and where the common tools fall short for agencies.
When a client asks how their AI search work is going, they want to see two things: that something changed, and that we did something about it. A tool that shows only the first half makes for a short, uncomfortable call. Our choice of reporting tool comes down to whether it can support both halves from one place.
We cover what a client report should say in what agencies should report about AI visibility. This article is about the software side: what we looked for, how the names clients keep mentioning compare, and the monthly report we build.
What we need from a tool
Our selection criteria are practical. We run several client programs at once, so we need a structure where each client is separate without separate bills. We need to show how AI engines treat the client's site, not only what they say about the brand. We need the output to leave the tool in a form the client already uses, which usually means Looker Studio. And we need to avoid rebuilding the report by hand each month, since that time is better spent on the fixes.
Mention counts satisfy none of that on their own. A count tells a client the brand appeared. It does not tell them whether the engines could read the new pages, which sources were cited instead of theirs, or whether any AI-referred visitor became a lead.
The platform we run client programs on
We use Promptwatch for client work, and reporting is a large part of why. Three of its features carry most of a report. Citation analytics shows which pages and domains engines cite, for the client and the competitors. Agent Analytics logs real AI crawler visits (ChatGPTBot, ClaudeBot, PerplexityBot and others) along with errors, so we can show whether new content was fetched. Visitor analytics tracks AI-referred traffic and conversions through a lightweight script or a Google Tag Manager template.
On the structure side, the agency plans are Kick-off at $199/mo, Growth at $399/mo and Scale at $799/mo. Each includes unlimited projects and prompts and 10 seats, so a client is a project, not a new subscription. Looker Studio and Google Search Console connectors feed the client-facing charts, and there is a REST API v2, an MCP server and Slack for the rest. We should be honest about one thing: white-label is available on Enterprise or custom arrangements, not on the self-serve plans. Our reports are branded in Looker Studio ourselves.
That combination is the deciding factor. The trackers can tell you the first half of the story. Crawler logs and traffic attribution tell you the second half.
The tools clients ask about
Clients often arrive with a name in mind. Here is how we describe each, using published information only.
Scrunch AI. It lists crawler traffic analytics, which is useful. The most common complaint in reviews, though, is weak reporting, with users exporting to Excel to build the client charts.
Profound. Its Starter plan is ChatGPT-only, and the broader features sit in an Enterprise tier from about $40k a year. It is sold as marketing agents more than as an AI-search reporting tool, so it rarely fits an agency's per-client economics.
Semrush AI Toolkit. The reporting is mature and clients recognize it. The cost is per domain: $99/mo each, with $99/mo for every extra domain, $99/mo for every extra user and $60/mo per 50 extra prompts. By the published figures a ten-domain agency pays about $1,000/mo, and the data stops at mentions without crawler or conversion evidence.
Otterly.AI. No PDF export. Looker Studio export begins at the Standard plan ($189/mo), it integrates with Semrush, and the data can be up to 7 days stale. Useful as a feed, thin as a report.
Peec AI. The Agency plan starts at $245/mo and includes a multi-client structure and white-label reporting, which is a real draw if white-label is a must. Each tier includes 3 models, and extra models are $35 to $165 per month each. It is a tracker, so the report covers visibility but not what the crawlers or the traffic did.
Our monthly client report
We keep the same outline every month so clients can compare periods.
- What we tracked. Prompts, engines, regions and competitors. Answers: what do these numbers cover?
- Visibility movement. Prompt trends and share of voice. Answers: are we showing up more in the questions that matter?
- Citations. Pages and domains engines used. Answers: whose content is being trusted, and where are the gaps?
- Crawler activity. Bots, pages fetched and errors. Answers: did the engines reach what we published?
- AI-referred visits and conversions. Answers: is any of this producing business?
- Work completed and next month. Pages fixed or published, with the plan. Answers: what did the retainer buy?
We do not put target numbers in a report before we have measured a client's own baseline, and we do not quote industry benchmarks we cannot source. A report with invented comparisons does more damage than a short one.
How we would start
If you are choosing a tool for the first time, set it up on one client for a full month and write the report above from it. Count the sections you could fill without leaving the tool. For us, that count is what separates platforms. You can try the agency plans through https://promptwatch.com?ref=101seomedia. For the wider workflow behind the reports, including how we connect agents to the data, see MCP in the agency workflow.