There's a growing blind spot in most digital marketing dashboards. Rank trackers report positions on Google for specified keywords. Google Analytics reports sessions, bounce rates, and goal completions. Google Search Console reports impressions and clicks. These tools have served the SEO world well for years — but none of them tell you whether your content is being cited by ChatGPT when a user asks a question relevant to your business.
That gap matters more every quarter. ChatGPT reached 400 million weekly users by early 2025. Perplexity processes hundreds of millions of queries monthly. Google's own Gemini-powered AI Overviews appear on a substantial share of all searches. If your measurement system only covers traditional Google rankings, you're flying blind on a meaningful and growing portion of where your potential customers are finding information.
This guide covers why traditional tracking fails for AI search, what manual testing looks like, which purpose-built tools are worth using, and how to build a measurement system that actually reflects your visibility across the current search landscape.
Why Your Rank Tracker Can't Measure AI Visibility
Rank trackers work by simulating a search query from a specific location, capturing the SERP that Google returns, and recording where your URL appears in the organic results. This is a solved technical problem that tools like Semrush, Ahrefs, and Moz have refined over many years.
The problem is structural: AI assistants don't return SERPs. When someone asks ChatGPT "which accounting software works best for small businesses in India?" the response is a paragraph of synthesised text, sometimes with citations, sometimes without. There's no rank position to record. There's no organic result placement. Your URL either gets cited or it doesn't, and traditional rank tracking has no mechanism to detect either outcome.
Google AI Overviews do appear on the SERP, so theoretically a rank tracker could detect them — but most commercial rank trackers haven't implemented this yet, or provide only limited coverage. They might detect whether an AI Overview is present on a tracked query, but not whether your specific domain is cited within it. That distinction matters enormously.
Perplexity, Claude, Gemini in standalone mode, Bing Copilot — each of these operates independently, cites different sources, and updates its knowledge through different mechanisms. A tool that tracks your position on Google.com captures none of this. The rank tracker was built for a world that's changing faster than the tools have adapted.
Manual Testing: Querying AI Systems Directly
Manual testing is the most accessible starting point, and it's more valuable than it sounds when done systematically. The core process is simple: assemble a list of 20 to 30 queries that represent how your target customers ask questions related to your services, then run those queries through each major AI system and record the results.
Setting Up a Query Testing List
Good test queries mirror natural language, not keyword strings. Instead of "SEO services Kerala 2026," use "who are the best SEO consultants in Kerala" or "how do I choose an SEO agency in Kochi." Include both branded queries (your name, your business name) and category queries (your service area without your name). Branded queries test whether AI systems know who you are; category queries test whether you appear among the recommended options in your space.
Testing Across Platforms
Run each query in ChatGPT (with web browsing enabled), Perplexity, Google's AI Overview (search in Chrome with a fresh session), and Gemini. Note: whether each platform produces a citation-style response, whether your domain appears in any citations, and what framing the AI uses when describing your area of expertise. These observations won't be quantified like a rank position, but patterns across months reveal whether your content efforts are building AI visibility.
Logging Results Consistently
Keep a spreadsheet with columns for query, platform, date tested, whether your site was cited (yes/no), and any notable phrasing in the AI's response that references your content. This log becomes a baseline. After two or three months, you'll see which queries consistently produce citations and which don't — and you can investigate the content differences between those two groups.
Tools Built for AI Visibility Monitoring
Several specialised platforms have emerged to automate what manual testing does manually, at scale and with historical tracking. These tools are still maturing — the space is less than two years old — but the leading options provide genuine value.
Otterly.ai
Otterly monitors your brand and competitor mentions across ChatGPT, Perplexity, Gemini, and Bing Copilot. You set up a brand profile and define queries to track, and the tool runs those queries on a schedule, alerting you when your domain appears or disappears from AI-generated responses. The dashboard shows citation frequency over time, which platforms cite you most, and how your share of voice compares to competitors on the same query set. It's one of the more mature options available as of early 2026.
Profound (formerly Search Pilot's AI tracker)
Profound focuses specifically on AI Overview tracking within Google Search, complemented by LLM citation monitoring. It integrates with Google Search Console to cross-reference your impressions data with detected AI Overview appearances, making it easier to see which pages are driving SERP presence without click-through. Useful for businesses primarily concerned with Google's ecosystem rather than standalone AI assistants.
Semrush AI Toolkit
Semrush has been building AI visibility features into its core platform, including an AI Overview tracker within its Position Tracking module. For businesses already paying for Semrush, this is the easiest entry point. It tracks whether AI Overviews appear on your monitored keywords, though domain-level citation detection within those overviews is still developing. The advantage is integration with your existing keyword data and site audit workflows.
LLMrefs.com
LLMrefs is a lighter-weight, lower-cost option that shows you which websites are most frequently cited by major LLMs on specific topics. Enter a topic, and it returns the domains that appear most often in AI-generated responses about that subject. It's less customisable than Otterly or Profound but useful for quick competitive benchmarking — understanding who gets cited in your space before investing heavily in optimisation.
Indirect Signals: What Google Search Console Still Tells You
Google Search Console doesn't report AI Overview appearances or LLM citations, but it contains several signals that correlate with AI visibility in ways worth understanding.
Impression-to-Click Ratio Changes
When a page starts appearing in AI Overviews, its click-through rate often drops while impressions stay flat or rise. This is because the AI answer satisfies some portion of the searchers, reducing their need to click. If you notice a sudden decline in CTR for a page without a corresponding rank drop, investigate whether an AI Overview is now present for its key queries. This pattern is a reliable indirect signal of AI inclusion.
Query Coverage Breadth
Pages with high topical authority — the kind of pages that get cited in AI answers — tend to rank for a wider variety of queries than their targeted keywords. In Search Console, filter by a specific page and look at the full range of queries it receives impressions for. A page with strong AI visibility often shows impressions for dozens of query variations it was never explicitly optimised for. Expanding query breadth without new keyword targeting is a hallmark of entity and topical authority building.
Brand Query Volume
Filter Search Console queries by your brand name. Month-over-month growth in branded impressions can indicate that AI citations are driving name recognition — users see your name cited in an AI answer, then search for you specifically. This correlation isn't proven in every case, but brand query growth that isn't explained by direct marketing campaigns or PR coverage is worth investigating against your AI visibility testing logs.
Building a Practical AI Visibility Measurement System
Rather than trying to use a single tool that solves everything — which doesn't exist yet — build a layered system that combines data from multiple sources into a coherent picture.
Layer 1: Monthly Manual Testing
Maintain a list of 25 to 30 queries (10 branded, 15 category/competitor queries). Run these through ChatGPT, Perplexity, and Google AI Overviews on the first Monday of each month. Log results in a shared spreadsheet. This takes about 90 minutes per month and generates the most qualitative data — you actually read what the AI says about your domain and competitors.
Layer 2: Automated Monitoring Tool
Choose one tool — Otterly, Profound, or Semrush's AI features depending on your budget and existing stack — to run automated checks daily or weekly on a set of priority queries. Set up alerts for when your domain appears or disappears. Review the tool's dashboard monthly alongside your manual testing results.
Layer 3: Search Console Signals
Run a monthly report in Search Console comparing impressions, clicks, and CTR by page for your top content. Flag any pages where CTR declined more than 15% month-over-month without a rank change. Investigate those pages manually to confirm AI Overview presence. Also track your branded query volume as a month-over-month trend.
Layer 4: GA4 AI Traffic Channel
In Google Analytics 4, create a custom channel grouping that captures known AI referral sources. Sessions from "perplexity.ai," "chat.openai.com," and similar domains arrive as referral traffic. Create a segment for these and track session count and goal completions monthly. Even if volumes are currently low, establishing this baseline now means you'll have historical comparison data as AI referrals grow.
Frequently Asked Questions
Is there a free tool to check if ChatGPT cites my website?
The most accessible free method is manual prompting: open ChatGPT with web browsing enabled, ask queries relevant to your business, and see whether your site gets cited in the response. LLMrefs.com is a free tool that shows which URLs various LLMs have cited on specific topics. It has limited query depth on the free tier but is useful for a quick pulse check. Google Search Console is also free and gives you indirect signals through impressions on AI Overview-eligible queries — look for queries where impressions rise but clicks drop sharply.
How often should I check my AI visibility?
Monthly is the right cadence for most businesses. AI models update their knowledge bases and citation patterns gradually, so daily or weekly checks produce more noise than signal. Run your manual query tests and tool reports on the first week of each month. Note any queries where you appeared last month but not this month, and vice versa. These changes are worth investigating — a disappearance might mean a competitor published better content, while a new appearance confirms a recent content update worked.
What metrics should I report to clients about AI visibility?
A practical AI visibility dashboard for client reporting should include: citation count across AI platforms (how many queries your site appears in), share of voice on target queries (what percentage of tested queries include your domain), AI-referred traffic in GA4 (sessions from identified AI referrers), brand search volume trend (monthly branded query impressions in Search Console), and a comparison benchmark against two or three competitors on the same query set. Keep the report to five or six metrics maximum — too many numbers obscure the story.