Today we are launching Agent Analytics, a new product that shows you how AI systems and answer engines access and use the content on your website. As services like ChatGPT, Perplexity, Microsoft Copilot and Google AI Overviews reshape how people discover information, the tools most teams rely on cannot see any of that automated activity. Agent Analytics closes that gap.
Key takeaways
- Traditional analytics depends on client-side JavaScript and cookies. AI bots do not execute either, so they are invisible in Google Analytics.
- Agent Analytics works from server-side logs and verifies every AI user agent against the IP ranges each provider publishes.
- It tracks two kinds of AI traffic: indexing crawlers that build training data and on-demand retrieval bots used for RAG.
- You can submit new URLs directly to select AI search indices and attribute the human visits that follow an AI mention.
Why we built Agent Analytics
Nearly every web analytics product, including Google Analytics, works by executing a JavaScript snippet in the visitor's browser and setting a cookie. AI crawlers and answer-engine agents typically do neither. The result is a set of blind spots that matter more every month: which AI crawlers are visiting your site, how often they index it, which pages they prefer, and whether that AI activity eventually turns into human visitors.
As discovery shifts to AI-first platforms, brands need a new kind of visibility into their own websites. Agent Analytics was built to provide it.
How it works
Agent Analytics uses server-side tracking. Every incoming request is logged, and requests claiming to come from an AI user agent are cross-checked against the IP ranges published by each provider so that spoofed crawlers are filtered out. Integration is available for the major hosting and edge platforms, including Amazon Web Services, Azure, Vercel and Cloudflare.
We distinguish between two categories of AI crawler:
- Indexing bots that crawl the web to build the datasets large language models are trained on.
- On-demand retrieval bots that fetch a page at the moment a user asks a question, powering retrieval-augmented generation (RAG) in answer engines.
A real-time log view shows every AI visit within seconds of it happening, with the bot, the provider, the page requested and the response status.
Understand your AI footprint
Agent Analytics tracks visits from OpenAI, Anthropic, Perplexity, Meta and others, so you can see where your pages are being read and surfaced in AI results. From there you can identify your top-performing content across engines and understand which pages are being picked up for answers.
Alongside the data, the product makes technical recommendations for making your site easier for AI systems to read: server-side rendering best practices, reducing reliance on client-side JavaScript, and getting structured data right.
Submitting new content to AI Search
Instead of waiting for a crawler's next routine visit, Agent Analytics lets you submit URLs directly to select AI search indices. Newly published or updated pages are discovered faster, which matters when answer engines are fetching content on demand.
Closing the loop with human referrals
AI activity only matters if it eventually reaches people. Agent Analytics tracks the human visitors who arrive after clicking a link embedded in an AI answer, completing the attribution path from AI conversation, to citation, to website engagement. For the first time you can see the whole chain in one place.
Visibility in AI used to stop at "we were mentioned." Now you can follow a page from the crawl, to the answer, to the human who clicked through.
Get started
Agent Analytics is available today. If you would like to see how AI systems interact with your site and start optimizing for the generative internet, get in touch with our team or explore the Agent Analytics feature page.