Agent Analytics: See How AI Models Access Your Website

There's a new source of traffic accessing your website, sending referrals, and influencing conversions. It doesn't show up in Google Analytics. Most teams can't see it at all.

AI models—ChatGPT, Perplexity, Claude, and others—are accessing specific pages on your site right now. When users ask these AI assistants questions about your category, product, or use cases, the models retrieve information from your pages to generate answers. Some of those users then visit your site as AI referrals.

This activity is real. It's measurable. And until now, it's been invisible.

The Blind Spot in Your Analytics

If you check Google Analytics for AI referral traffic, you might see some. But what you're seeing is incomplete—and the most important part is missing entirely.

Google Analytics shows you when a human clicks through from an AI answer to your website. That's referral traffic, and it's valuable to track.

What it doesn't show you:

• Which pages AI models access to generate those answers
• Which AI models are accessing which pages
• How frequently different models access your content
• Which pages are accessed most by AI vs. never accessed
• Trends in AI access patterns over time

The access happens at the infrastructure level—AI models make server requests, retrieve content, and use it to inform responses. None of that triggers the JavaScript tracking that powers Google Analytics.

You see the outcome (referral clicks), but not the activity that creates that outcome (page access by AI models).

Why This Matters Now

Understanding which pages AI models access isn't academic. It has direct implications for how buyers discover and evaluate your product.

When a potential customer asks ChatGPT "what's the best solution for [your category]," the AI's response is shaped by which of your pages it accesses and what information it retrieves. If it accesses your homepage but not your use case pages, the answer reflects that limited view. If it accesses documentation but not positioning content, that shapes the response.

Different AI models access different pages. Perplexity might access your blog posts. Claude might access your documentation. ChatGPT might access comparison pages. Each model's behavior creates different representations of your product in AI-generated answers.

And all of this happens before the user ever clicks through to your site—if they click through at all. Many AI-assisted research sessions end without a website visit. The AI access is the entire interaction.

Introducing Agent Analytics

Agent Analytics is a free analytics product that provides visibility into AI model activity on your website.

It shows:

• Which pages AI models access — See exactly which URLs different AI models request
• Which AI models are accessing them — Break down activity by ChatGPT, Perplexity, Claude, and other models
• Referral traffic from AI models — Track when AI-assisted users visit your site
• Conversions tied to AI referrals — Measure conversion behavior for AI-driven traffic
• Trends over time — See how AI access patterns change
• Recent activity — View live and recent AI requests

All data comes from server-side requests captured at the CDN and infrastructure level. This isn't modeled, inferred, or estimated—it's direct observation of actual AI model requests hitting your infrastructure.

What This Looks Like in Practice

Here's what teams learn when they start seeing their AI traffic:

Page-Level Visibility

You can see which specific pages AI models access. Your pricing page might get frequent requests from Perplexity. Your API documentation might see heavy access from Claude. Your blog posts might be accessed by ChatGPT.

This is observable fact—these requests happened, at these times, from these models.

Model-Level Differences

Different AI models exhibit different access patterns. Some models access more pages per session. Some focus on specific content types. Some return to the same pages repeatedly.

Agent Analytics shows these differences so you can see how each model interacts with your site.

Referral Behavior

When users click through from AI-generated answers, Agent Analytics tracks that referral traffic alongside the original AI access. You can see which pages were accessed by the AI, then which pages the user visited after clicking through.

Conversion Attribution

If your site tracks conversions, Agent Analytics connects those conversions back to AI referral sources. You can measure how AI-driven traffic converts compared to other channels.

What This Is (and Isn't)

Agent Analytics is an analytics and visibility product. It shows you what's happening with AI traffic on your site.

It is not:

• An SEO tool or ranking service
• An optimization platform
• A tool that tells you "what AI thinks" about your brand
• A predictive or inferential system
• A competitive intelligence tool

We don't claim to know why AI models access certain pages, what topics they "care about," or how to "improve your AI ranking." Those would be speculation.

What we do provide is factual data: these models made these requests to these pages at these times. That's observable reality, not interpretation.

Who This Is For

Agent Analytics is built for teams who care about attribution, traffic visibility, and understanding how different channels drive business outcomes.

If you're in marketing, growth, RevOps, or analytics, and you want to see how AI models interact with your content—not because you're chasing rankings, but because you want complete visibility into your traffic sources—Agent Analytics gives you that view.

This is especially relevant if:

• You're seeing AI referral traffic in Google Analytics and want to understand what drives it
• You suspect AI models are accessing your content but have no way to confirm it
• You want to measure AI as a channel alongside organic, paid, and other sources
• You need to report on AI traffic and conversions to stakeholders

Why We Built This

At Salespeak, we help B2B companies engage buyers through AI. In working with customers, we kept hearing the same question: "How do we know if AI models are even accessing our content?"

Standard analytics tools couldn't answer that question. The data exists—it's in server logs and CDN access records—but it wasn't surfaced in a way that made it actionable for marketing and growth teams.

Agent Analytics makes that data visible. It's a missing layer of observability for a channel that's already influencing how buyers discover and evaluate products.

Getting Started

Agent Analytics is free. It takes minutes to set up—you add a tracking snippet to your site, and data starts flowing within 24 hours.

You'll see:

• A dashboard showing AI model activity across your site
• Page-level breakdowns of which AI models access which content
• Referral traffic and conversion data tied to AI sources
• Trend lines showing how AI access patterns change over time

This is visibility into a traffic source that already exists. AI models are already accessing your site. Agent Analytics just makes that activity visible.

See your AI traffic with Agent Analytics