A Hands-On Review and Setup Guide for GEO/AEO Tracking

You can rank #1 on Google and still be invisible on ChatGPT. That gap — between classic SEO rankings and how AI engines describe your brand — is exactly what Rank Math’s new AI Visibility module is trying to close. It’s tucked into the Rank Math SEO dashboard under a shiny “New!” tag, sitting right next to Titles & Meta and Schema Templates.

I went through the module screen by screen — setup, scoring, competitor tracking, and raw transcripts — to see whether it’s a genuinely useful tool or just a dashboard widget dressed up in AI branding. Here’s the breakdown.

What “AI Visibility” Actually Measures

Once you open the tab, you land on a dashboard with three sub-tabs: Dashboard, Analyses & Transcripts, and Reports & Export. The main dashboard gives you three headline numbers:

• Global AI Visibility Score — an aggregate score (out of 100) across every brand/product you’re tracking
• Analyses in the last 24h — how many fresh AI queries have run
• Average rank position across tracked brands

Below that sits a brand table with columns for AI Visibility Score, Avg Rank, Avg Sentiment, Mentions, Citations, and last-updated date. In the test account, four brands were being tracked — Metricool (69/100), My Hours (0/100, freshly added), Wix (62/100), and Zara (38/100) — which tells you the score isn’t a vanity number; it clearly separates brands with strong AI presence from ones that are effectively invisible.

 VERIFIED   The scoring model combines rank position, sentiment, mention count, and citation count into a single composite score — this is directly visible in the column structure of the dashboard table.

 VENDOR CLAIM   Whether Rank Math’s scoring weights these four inputs identically across all customers, or whether the algorithm adjusts per industry, isn’t disclosed anywhere in the UI. Treat the 0–100 score as directional, not a precise benchmark.

Setting Up a Brand: The “Add Brand / Product” Flow

Adding something to track is a single modal with five fields:

• Brand/Product Name
• Brand/Product URL
• How would you describe your brand/product? — a free-text box that’s almost certainly fed to the AI model as context for how to identify and represent your brand correctly
• Target Country (optional)
• Frequency of Analyses — defaults to Weekly

Then comes the part worth paying attention to: AI Platforms. At the time of this review, only ChatGPT is live. Gemini, Perplexity, Claude, and Google AI Overview are all greyed out with “Coming Soon” labels.

 DISPUTED   This is the single biggest caveat of the whole feature. If your actual customers are asking Perplexity or Gemini about your category — which, depending on your niche, could be a huge chunk of AI-driven discovery — this tool currently tells you nothing about that. For now, read every score in this dashboard as a “ChatGPT Visibility Score,” not a general AI visibility score.

Brand Detail Page: Where the Real Value Is

Clicking into a tracked brand (I used the time-tracking tool “My Hours” as the test case) opens a detail view with four tabs: Overview, Queries, Competitors, and Raw data / Transcripts.

• Overview — AI Visibility Score, Recent Mentions, Avg Sentiment, and a Top Competitor card. In the My Hours example, it surfaced “Harvest” automatically, without the user entering a competitor name.
• Competitors — a table of competing brands, their average sentiment, and mention counts across the same query set
• Queries — auto-generates a baseline set of natural-language prompts a real buyer might type into ChatGPT, e.g. “best time tracking tool for freelancers” or “I need a budget manager that also tracks hours.” Each query can be toggled on/off, with a “Regenerate baseline queries” button if the first batch doesn’t match your buyer language
• Raw data / Transcripts — the most transparent part of the tool. Every query run gets a full transcript: the original prompt, the model’s complete response, and a status flag (“Partial” appeared frequently, meaning the brand was mentioned but not featured as a top pick)

In the example transcript, ChatGPT’s answer to “best time tracking tool for freelancers” listed Toggl Track, Harvest, Clockify, Timely, Bonsai, Paymo/Everhour, and Hubstaff — with My Hours notably absent from the “Fast picks” list.

This transcript layer is what turns the feature from “another dashboard score” into something a GEO practitioner can actually act on.

You’re not just told your score is low — you can see exactly which competitors ChatGPT recommends instead, and under what framing, which tells you what content gaps or trust signals you need to close.

The Good, the Gaps, and Who This Is For

What works

• Native integration inside a plugin most WordPress sites already run — no separate n8n workflow or third-party AI-tracking subscription needed for a baseline check
• Auto-generated query sets save the manual work of guessing buyer-intent prompts
• Full transcript visibility (not just a score) is the difference between a vanity metric and a diagnostic tool
• Automatic competitor detection — most GEO tools require you to manually name competitors

What’s missing (for now)

• Single-platform coverage (ChatGPT only) means “AI Visibility Score” currently overstates its own scope
• Weekly analysis frequency may be too slow if you’re actively testing content changes and want faster feedback loops
• No visible methodology for how sentiment percentage or the composite score is calculated

Who should use it today

Agencies and site owners already inside the Rank Math ecosystem who want a first-pass, ChatGPT-specific visibility audit without adding another tool to the stack. If you’re running deeper multi-model GEO tracking (Perplexity, Gemini, AI Overviews) for client reporting, you’ll still need a supplementary setup until Rank Math ships the other platforms — Reports & Export is there for handing a client-facing summary once that data matures.