Where does the data in Semrush’s AI Visibility Toolkit come from?

Semrush’s AI Visibility Toolkit uses several data sources to give you accurate insights into how your brand and competitors show up in AI-generated answers. Each report pulls from a slightly different database and update schedule.

For over 17 years, Semrush has been the industry leader in SEO data. Now, we’re bringing the same expertise and scale to AI search, giving you trusted insights into how your brand appears in AI-generated answers.

Prompt Database (Visibility Overview, Competitor Research, Prompt Research)

These AI Analysis reports are powered by Semrush’s AI prompt database—a collection of over 100 million prompts and responses on ChatGPT, Google AI Overviews, and AI Mode.

• Data is updated monthly
• Currently covers 6 regional databases (US, UK, Canada, Australia, India, Spain)
• Prompt responses are captured from real requests and not via any APIs of LLMs
• Reports can be run up to 300 times per day
• Coverage will expand as more AI platforms and regional databases are added

How Does Semrush Get Its Prompt Data?

We source billions of real prompts from AI search clickstream data and Google’s keyword dataset for AI Overviews. These are organized into meaningful Topics, with duplicates removed and phrasing simplified—while always preserving the original intent and semantics.

How Does Semrush Calculate AI Volume?

Unlike keywords, individual prompts are often too specific and unique to measure directly. That’s why Semrush calculates topic-level volume in our Prompt Research report.

Each topic groups together related prompts that move in the same “semantic direction,” giving you a meaningful estimate of demand.

To estimate topic volume, we combine third-party data on real AI interactions with Semrush’s machine learning models. 

This approach allows us to provide a reliable, data-driven view of how often people engage with different topics across AI platforms—without relying on raw, one-off prompts.

How Does Semrush Calculate AI Topic Difficulty?

Topic Difficulty shows how challenging it is for your brand to get mentioned in AI answers for a given topic. It’s calculated by looking at two main factors:

1. Competitor strength. If the brands most often mentioned for this topic are already well-known and authoritative, it’s harder to break in.
2. Opportunity size. We compare how many positions are available for this topic versus the average across all topics. If there are fewer opportunities to be cited for this topic compared to others, it becomes harder to gain visibility

In short: the more established the competing brands and the more limited the available mentions, the higher the difficulty score.

How Does Semrush Extract Brands from LLM Responses?

To measure brand mentions in AI answers, the AI Visibility Toolkit is able to identify brands based on their names and context mentioned within AI-generated answers. 

To identify brands, Semrush uses its advanced proprietary AI brand extraction system. Unlike a simple text match, this system understands context and sub-brands or products of a main brand—so it can tell the difference between Tesla (the EV company), Nikola Tesla (the scientist), and Nikola Tesla Airport in Belgrade.

This means your brand gets recognized even if users type it with different spellings or variations.

As our technology is continuously improving, we’re actively enhancing our database to ensure accuracy and inclusivity for brands of all sizes.

Which ChatGPT Model Does Semrush Use to Collect AI Visibility Data?

The AI Analysis prompt database in the Semrush AI Visibility Toolkit analyzes responses generated by the latest model of ChatGPT in search mode. This is the version of ChatGPT that the toolkit monitors to collect prompts, mentions, and citations for visibility reporting.

In the Brand Performance database, which provides sentiment analysis on your brand, Semrush uses the latest ChatGPT model both with and without search mode (aka SearchGPT) enabled. You can filter your Brand Performance reports based on ChatGPT or SearchGPT.

Brand Performance Database

Brand Performance reports allow you to closely monitor the narrative around your brand, as it is represented in AI answers. 

To create these reports, we maintain a large repository of queries submitted to various AI platforms and use it to identify both branded and non-branded queries that directly reference or can be contextually associated with your domain. This identification process is powered by our proprietary technology.

This data focuses on how AI platforms describe your brand, including sentiment and share of voice.

• Data is updated weekly
• Covers multiple platforms including Google AI Mode, ChatGPT, SearchGPT, Perplexity, and Gemini

Prompt Tracking (Position Tracking integration)

Prompt Tracking monitors your visibility for specific prompts in Google AI Mode, AI Overviews, and ChatGPT Search. 

This data is collected by Semrush’s Position Tracking tool querying your target prompts on your target platform (and location) each day, the same way Position Tracking collects data on traditional Google search positions. 

AI Search Site Audit

Site Audit data comes from Semrush’s own crawler, combined with checks against eight different AI crawlers (such as OAI-SearchBot). 

This data:

• Updates every time you run a crawl
• Shows whether your site is accessible and optimized for AI engines

Summary of AI Search Data:

• Prompt data: 100M+ prompts, volume and difficulty for topics, monthly updates
• Brand performance data: separate database, sentiment and share of voice, weekly updates
• Prompt Tracking: daily updates on custom prompts 
• Site Audit: updates on each crawl

This structure ensures you’re getting both breadth (millions of prompts across AI platforms) and freshness (daily and weekly updates where they matter most). 

Why AI Search Data Matters

AI search and LLM responses are fast-changing and highly personalized, which means no platform can provide exact numbers on visibility.

Semrush invests heavily in data — with a database of over 180 million prompts and the resources of a company trusted by 10M+ marketers worldwide. Our AI visibility metrics are built to give you reliable directional signals you can use to spot trends, benchmark competitors, and guide strategy with confidence.