Issue 13 · Nov 2–8, 2026Editions · Glossary · Archive · Method · RSS · HU / ENG

News of the AI era, curated weekly

Glossary › Search and marketing

AI-forgalom

in English: AI Traffic

basicspreading

AI traffic refers to website visits originating from artificial intelligence: humans clicking through from an AI response, as well as AI bots and agents reading the site.

In plain terms

In the past, people mostly reached a website by typing its address, clicking a Google result, or following a link from another site or from social media. Today, more and more people ask ChatGPT, Gemini, Perplexity or Copilot, and arrive through the link in the answer. This is one half of AI traffic.

In the other half, the visitor is not a human but a machine. AI bots read through the page to learn from it, or to answer someone's question right then. Increasingly, agents also arrive, searching, comparing or buying on a person's behalf.

Imagine you have a shop. Until now, customers entered from the street based on the window display and the company sign. Now a consultant has appeared who recommends you at the other end of the city, or perhaps does not. Sometimes they send a customer to you, sometimes they come in to look around so they have something to talk about. AI traffic is these two types of visitors together.

It is worth paying special attention to it because it behaves differently than traffic from traditional search engines. It usually means fewer visitors, but they often arrive more targeted, as they have already received a preliminary answer to their question. Whether this leads to more sales varies by page, and measurements are not consistent. Meanwhile, many people do not click through at all, as they are satisfied with the AI summary.

How it works

  1. AI company robots (crawlers) regularly visit websites. Some collect text to train models, while others build search indexes.
  2. When someone asks an AI assistant a question, the system often queries several pages live to provide an up-to-date answer. This is also a machine visit, but it is triggered by a specific person's question.
  3. In the response, the AI names sources and adds links. If the user clicks, they arrive at you as a human visitor. Web analytics often shows where they came from (for example, chatgpt.com or perplexity.ai), but not always: the source of visitors coming from mobile apps is often lost.
  4. Agents go one step further: instead of humans, they browse, fill out forms, and compare prices. Your site can thus reach a customer even if they never see it themselves.
  5. The website owner can sort this traffic from server logs and analytics. Using a robots.txt file or settings from the hosting and network provider, they can decide which AI robots are allowed in.

An everyday example

You ask ChatGPT which guesthouse on Lake Balaton is dog-friendly and has its own jetty, then click the link below the answer and land on the guesthouse's website. The owner sees in their analytics that you arrived from chatgpt.com: you are their AI traffic.

Who is affected

marketerYou must measure and handle a new traffic channel separately, while clicks from traditional search decline in many places. → Marketing
entrepreneur, executiveYour revenue may depend on whether AI assistants mention your company and send inquiries to you.
developerThey must configure which AI robots can access the site and monitor how much load they place on the server. → Code
everyoneAnyone asking AI questions also generates AI traffic, thereby shaping which sites remain viable on the web.

A common misconception

MythAI traffic is simply bot traffic that must be filtered out or blocked.
In realityPart of AI traffic consists of real human visitors who clicked through from an AI answer, and they often arrive with serious intent. It is also worth deciding carefully about blocking bots, because they are not all the same: blocking a crawler that collects data for training generally does not exclude you from AI search results, whereas blocking the crawlers used for search and live answering does. If you exclude those too, AI is less likely to cite you and will, at most, mention you based on what others have written about you.

Limits and risks

Similar, but not the same

Related concepts

Broader

Webforgalom, Webanalitika, Digital marketing

Narrower

AI referral traffic, AI-robot (crawler), Agentforgalom

On the edge

Generative search engine optimization, AI visibility, zero-click search, AI overview, robots.txt

A brief history

The term began spreading after ChatGPT launched in November 2022, when website statistics started showing visitors arriving from AI chatbots and the data-collecting bots of AI companies. It became a real marketing topic from 2024, when Google launched AI Overviews and OpenAI introduced the search function in ChatGPT.

Where things stand

At the end of 2026, AI traffic on most websites is still a small share of total traffic, but it is growing quickly, and analytics tools increasingly show it as a separate channel. Measurement methods and the rules for letting bots in are not yet uniform.

Status: 2026 November

In the news

Appears in 5 Agent Age stories; the curve shows weekly occurrence in the news magazine.

Also known as

AI-forgalom · MI-forgalom · AI traffic · AI referral traffic · AI referral traffic · LLM traffic · AI bot traffic · agent traffic

Sources

This entry was machine-edited and has undergone machine expert review; the “In the news” section updates automatically from Agent Age stories.