AI Search Visibility Agency Dashboard for 15 Clients: Platform for AI Overviews, ChatGPT, and Perplexity Tracking

People search for an "AI search visibility agency dashboard for 15 clients" because the math on brand plans falls apart at a portfolio. Fifteen separate Essential seats is fifteen invoices and fifteen logins. The surfaces they want in the report are Google AI Overviews, ChatGPT, and Perplexity. Gemini and Claude show up the week after.

This post is about that dashboard shape. It is not a headcount claim. We will not say 1001 SEO Media "has 15 clients." Some months the GEO roster is smaller. Some months it is larger. The software problem is the same once you pass a handful of brands: unlimited projects, a seat for each lead, and one way to pull Overviews, ChatGPT, and Perplexity without a custom export per domain.

The platform we use is Promptwatch. Agency Kick-off is $199/mo: unlimited projects, unlimited prompts, 10 seats, 10,000 responses, 10M crawler logs, 7-day trial. Growth is $399/mo (25,000 responses). Scale is $799/mo (65,000 responses). That is the dashboard. Our services are the humans who read it and change pages.

Why brand plans break on a roster

Essential is $95/mo for one project and 50 prompts. Correct for a single company running its own GEO. Wrong as the agency pattern. Professional ($245/mo, 2 projects) and Business ($579/mo, 5 projects) still cap projects. At a 15-brand thought experiment you are either stacking Business-class invoices or you are on an agency SKU.

Kick-off's unlimited projects exist for that thought experiment. Ten seats cover the agency operators plus a login for the client who wants to look. We do not put every brand on 200 prompts on day one. We put each brand on a tight list and let unlimited prompts mean we are not rationing rows across the roster.

Explore (free, ChatGPT-only, 10 prompts) is how we show one stakeholder the UI. It cannot be the 15-brand dashboard. It cannot see Perplexity or AI Overviews.

What the dashboard has to show

Per client project, every week:

AI Overviews / AI Mode. Prompt-level presence. Search Console's generative AI performance reports stay on the client's Google account for official impressions. Promptwatch is the prompt layer GSC does not store.
ChatGPT. Mention vs citation vs competitor URL.
Perplexity. Footnote domain, not just a vibe that "we were in the answer."

Paid Promptwatch engines include those plus Gemini and Claude. We still lead the client PDF with Overviews, ChatGPT, and Perplexity because that is what the search asked for. The extra engines go in an appendix so a Gemini-only miss does not hijack the meeting.

Shared across the roster:

Crawler logs, because one client's "block AI bots" CDN preset will otherwise look like a content failure.
Visitor analytics (script or GTM) so we do not report mentions as sessions.
A review inbox for Content Agents when a project is on Webflow or Framer. Most clients are not. Most drafts stay in their CMS.

We do not roll 15 brands into one vanity "network AI score." Each project keeps its own prompt list. A dentist and a B2B vendor do not share a denominator.

How we report without inventing a score

The weekly artifact is boring on purpose:

Prompts that flipped from cited to absent (or the reverse) on more than one engine.
Crawl errors that would explain a flip.
One recommended URL change, or one third-party mention to earn.
AI-referred sessions only if the pixel is live.

That is enough for a retainer meeting. It is also all we can defend. We will not attach fake lift percentages to a dashboard export.

If you are building this in-house for ~15 brands, buy Kick-off (or Growth if response volume will blow past 10,000). Put each brand in its own project. Give the client a seat. Do not build a Notion database of ChatGPT pastes.

Profound Starter ($99/mo annual, ChatGPT-only) does not cover Overviews and Perplexity. Otterly Lite ($29/mo) plus add-ons and a weekly-ish refresh is a painful way to staff a roster. Peec at $95/mo is one brand. This is why the agency SKU exists.

The work the dashboard cannot do

Someone still has to unblock OAI-SearchBot, write a sentence a model can lift, and pitch the listicle the fan-out already retrieves. We do that as the agency. Promptwatch will not sit on the client's CMS. Kick-off will not write the monthly narrative.

A 15-client dashboard without that labor is a slide deck. With it, it is a GEO practice. We described the single-brand version in how we use Promptwatch. Scale the project count. Do not scale the excuses.

FAQ

Do you have exactly 15 GEO clients?

That number is the search query, not our roster. Ask us for availability on a call. We will not invent a headcount for this URL.

Is Kick-off enough for fifteen busy brands?

Unlimited projects and prompts, yes. 10,000 responses and 10 seats can get tight if every brand is on a huge daily pull. Then Growth ($399) or Scale ($799). We start on Kick-off and watch response usage.

Can we white-label the dashboard?

White-label is on Promptwatch Enterprise/custom, not on Kick-off. On Kick-off we export into the client's existing report template and say the data source out loud.

What to do this week

Count the brands that actually need Overviews + ChatGPT + Perplexity tracking. If that number is more than two, look at Kick-off ($199/mo), not stacked Essentials.
Create one Promptwatch project per brand. Same prompt-writing rules: sales language, not a keyword dump.
Install visitor analytics only on properties you own. Confirm crawlers per domain.
Write the weekly template (flips, crawls, one action) before you add more prompts.
If you want us to operate the dashboard and the tickets, email hello@1001seomedia.com. The platform is Promptwatch. Fifteen is a capacity question, not a slogan.