---
title: "Half the traffic on your website might not be human..."
url: https://sam-page.com/thinking/half-the-traffic-on-your-website-might-not-be-human
date: 2026-04-02
author: Sam Page
type: Post or article
categories: [Emerging tech and AI]
reading_time: 2 min
summary: "Which isn't necessarily a bad thing."
first_published: https://www.linkedin.com/feed/update/urn%3Ali%3Ashare%3A7445502559559204865
---

# Half the traffic on your website might not be human...

Which isn't necessarily a bad thing.

![](https://sam-page.com/img/thinking/half-the-traffic-on-your-website-might-not-be-human-li1.webp)

Human Security's 2026 State of AI Traffic report found that automated traffic is growing 8x faster than human traffic. AI agent traffic alone grew 8,000% in 2025.

The instinct is to see that as a problem, but I think it is more nuanced than that.

Some of that bot traffic is genuinely useful. AI agents are browsing, researching, and buying on behalf of real people.

That's not fake activity. It's human intent, delivered by a machine.

And the analytics on that traffic matters now too.

If an AI agent is visiting your site to compare pricing for a real customer, that's a lead. If it's pulling your content into a recommendation for someone, that's visibility.

Filtering all of that out means you're ignoring a growing chunk of how people actually find and evaluate businesses.

The problem is that some analytics platforms can't tell the difference, or businesses aren't looking at both.

A scraper hitting your site looks the same as an AI agent genuinely shopping for someone.

HUMAN Security found that only 0.5% separates benign automation from malicious automation. That's how thin the line is.

So the question isn't "how do we block the bots?"

It's "do our tools understand who's actually on the other side?"

The definition of a website visitor has changed.  
Most of the tools we use to measure them haven't.

How confident are you that your analytics are built for this?