Report on LLM and AI-generated traffic

This use case article explores how to use the Customer Journey Analytics derived fields capability as a foundation to report on LLM (Large Language Model) and AI-generated traffic.

Detection methods

To detect LLM and AI-generated traffic, distinguish between:

Three common core detection methods to identify and monitor LLM and AI-generated traffic are:

The following table illustrates how the detection methods can be used against different LLM and AI interaction scenarios.

Challenges

LLM & AI agents demonstrate complex and evolving behaviors when interacting with digital properties. These technologies operate inconsistently across platforms and versions. This inconsistency creates unique challenges for data professionals. The behavioral patterns vary significantly and depend on the specific AI platform, version, and interaction mode used. This operational diversity complicates efforts to track and categorize LLM and AI-generated traffic within standard analytics frameworks. The complex nature of these interactions, combined with their rapid evolution, requires nuanced detection and classification methods to maintain data integrity:

Detection signatures

As of August 2025, the following specific signals can be identified for each of the detection methods.

User agent identification

Referrer classification

Query parameter detection

Implementation

You can report on LLM and AI-generated traffic within a typical Customer Journey Analytics setup (connection, data views, and workspace projects) through the specific setup and configuration of derived fields, segments, and workspace projects.

Derived fields

To configure detection methods and detection signals use derived fields as the foundation. For example, define derived fields for user agent identification, query parameter detection, and referrer classification.

LLM/AI user agent identification

Use the Case When derived field functions to define a derived field that identifies LLM/AI user agents.

{modal="regular"}

LLM/AI query parameter detection

Use the URL Parse and Classify derived field functions to define a derived field that detects query parameters.

{modal="regular"}

LLM/AI referrer classification

Use the URL Parse and Classify derived field functions to define a derived field that classifies referrers.

{modal="regular"}

Segments

Set up dedicated segments that help you to identify events, sessions or people related to LLM and AI-generated traffic. For example, use the derived fields that you created earlier to define a segment that identifies LLM and AI-generated traffic.

{modal="regular"}

Workspace project

Use the derived fields and segments to report and analyze on LLM and AI-generated traffic. For example, see the annotated project below.

{modal="regular"}