How to Track and Monitor Traffic & Conversions From AI Search in 2025

AI search attribution means tracking when a user discovers your brand or converts after using an AI tool like ChatGPT, Gemini or Bing Copilot.

Tracking performance in AI search is tough because most users who get their answers from ChatGPT, Gemini, Perplexity or other LLMs rarely click through to your website. In fact, only about 1% of users actually visit a site from an AI tool. 

Most AI search experiences are suited for discovery, not direct traffic. Results are personalised and can change day by day. There’s no official analytics dashboard for AI traffic, so website owners are left in the dark.

You can track some AI-driven visits by setting up filters in Google Analytics 4. GA4 filters help you spot sessions from AI platforms if the traffic includes referrer data.

You can use a regex filter in your User Acquisition report to search for sources like chatgpt.com, perplexity.ai, copilot.microsoft.com or gemini.google.com. The filter won’t catch every visit, but it helps you see traffic spikes from AI search and which content is attracting it.

You can also experiment with AI tracking tools for a general sense of brand visibility, not for attribution.

These tools offer AI visibility features like tracking brand mentions in AI-generated answer boxes (like AI Overviews, ChatGPT responses, Gemini summaries) and surfacing competitive insights.

• Integrates with existing SEO workflows
• Offers competitor benchmarking and visibility tracking within one familiar dashboard
• Useful for marketers already using these platforms

• Priced per domain or AI model so costs can add up quickly
• Only tracks keywords within its own database, so coverage depends on your brand size and niche - big brands with broad topics will see more mentions, but niche players may see limited results
• Doesn’t track traffic, conversions or user sources directly, only visibility data
• Lacks prompt-level tracking or insights on user sentiment

This lets you query AI models (like GPT-4, Gemini or Claude) directly via API for branded or competitor search terms, so you can manually monitor how your brand appears in AI-generated answers.

• Full control over what keywords, questions or prompts you track
• No dependency on third-party dashboards or pre-set databases
• Flexible: Can be adapted for niche brands or to track very specific prompts

• Requires technical setup and API management skills
• No automated dashboard or alerts - results need manual analysis and tracking
• API usage costs can add up, especially if tracking a lot of queries or running frequent checks
• Lacks built-in brand safety features (unless you build them yourself)

This method can be time-consuming and hard to scale without additional tooling

A full-featured AI visibility platform built for Generative Engine Optimisation (GEO), Profound tracks mentions on ChatGPT, Google AI Overviews, Gemini, Claude and more, with detailed dashboards, sentiment insights and many other features.

• Covers multiple AI models, languages and regions
• Offers enterprise-grade compliance

• Not for everyone - High cost (starting around $499/month) makes it better suited to larger teams or enterprises
• Complex interface with a steep learning curve for non-technical users

Promptmonitor offers dual tracking: AI bot crawling analytics (e.g., whether ChatGPT or others are accessing your site) and visibility tracking (how often you’re mentioned/referenced).

• Tracks multiple AI models (ChatGPT, Gemini, Claude, etc.)
• Real-time insights into bot behaviour and brand mentions 
• Supports local tracking
• Great fit for SMEs and marketing teams with scalable pricing

• Usage-based pricing (credit model) may become expensive with heavier usage
• Newer tool - not as established or enterprise-tested yet

1. Navigate to the User Acquisition Report: Go to Reports > Acquisition > User acquisition.
2. Edit the Report: Click the pencil/edit icon in the top-right corner.
3. Add a Filter for AI Sources:
   • • Under Filters, click Add filter.
     • Set the Dimension to Session source and choose Matches regex (ignore case).
     • Use this regex: chatgpt\.com|perplexity\.ai|chat\.openai\.com|copilot\.microsoft\.com|gemini\.google\.com|blackbox\.ai|julius\.ai
     • Apply the filter to show only traffic from these AI tools.

Optional

• For a breakdown of traffic, ensure Session source is the primary dimension.
• To view landing page data, click + Add dimension, select Landing page and adjust the report.

1. Visit the SEMrush Sensor.
2. Under SERP Features Occurrence, select AI Overviews and adjust Date, Industry, Location or Device as needed
3. For individual terms:

• • • Go to Organic Research.
    • Enter the domain, then click View All Keywords or Positions.
    • In the filter bar, click SERP Features > SERP Features on SERP > AI Overview.
    • Add additional filters (e.g., intent).

Note: This is not 100% accurate but provides a good overview.

Rocket uses a combination of dedicated dashboards, industry tools and custom tracking to monitor AI search performance for every client.

We set up tailored dashboards for each client to track a range of AI visibility metrics, ensuring every account has a clear, up-to-date view of their performance in AI search environments.

For top-line numbers, such as the number of AI overviews or brand mentions a client receives, we leverage Semrush and Ahrefs. These tools help us benchmark progress, monitor competitors, and provide consistent reporting on AI-generated search visibility.

To dig deeper, we’ve built an in-house tool that tracks specific key LLM prompts for each client. This lets us see exactly how our clients show up for targeted queries and monitor changes in real time, beyond what third-party dashboards can provide.

• Update all key forms (contact, quote, lead, newsletter sign-up) to include a required or optional field such as:
  • How did you hear about us?
    • Google/Search
    • Social Media
    • Referral
    • ChatGPT/AI Search
    • Other (please specify)
• Train your team: Make sure sales/lead handlers recognise and record when a prospect mentions AI, even in conversation or emails.

Check your analytics (GA4, HubSpot, etc.) for unexplained spikes in “direct” traffic. Map these spikes to dates when your brand appears in AI summaries, ChatGPT outputs, or Gemini answers for core keywords.

• Each week, ask ChatGPT, Gemini and Bing Copilot your top commercial and brand queries (“Best home stager Sydney,” “Who are the top gym franchises in Australia?”). Document where (if at all) your brand appears - take screenshots or notes.
• Cross-check with lead spikes - When you see a bump in inbound interest, compare with the weeks your brand was visible in AI answers.

Use a spreadsheet to log queries, dates and outcomes for easy trend-spotting over time.

There’s no single, foolproof way to track all AI-driven discovery and leads and it’s important to understand the current limitations:

• Low CTRs: AI-generated answers often provide instant, in-line information and many users will get what they need without ever clicking through to your site.
  
• Highly personalised results: Results for a query are based on each user’s prompt and their personal search history. This means the same query can produce different answers for different people, or even for the same person at different times.
  
• Very volatile: The content of AI answers can shift day-to-day as models update or as the internet’s data changes. Even running the same search back-to-back can give slightly different outcomes.
  
• Multi-channel journeys: Many users discover brands through AI search, but will ultimately visit your website via traditional channels (like Google Search, direct URL entry or social media) when they’re ready to convert or get in touch.

 A: Not yet. Even the best tools only show mentions or rankings, not user-level attribution.

 A: These tools help monitor visibility, but don’t fully solve AI attribution. For most marketers, start with manual tracking and analytics review.

A: Yes. More platforms are working on attribution, and Rocket will update our methods as tools improve.

A: Add “AI/ChatGPT” as a source on your forms, and review analytics for direct traffic patterns.

AI and LLM search are rewriting the rules of digital marketing. While attribution is still catching up, the brands that start adapting now will be first to capture new leads as search evolves.