AI Conversation Fragments Appear in Google Search Console Reports

Image credit: Search Engine Journal

Fragments of artificial intelligence conversations are appearing in Google Search Console performance reports, offering new insights into user interaction with AI Overviews and AI Mode since their August 11 rollout.

This data provides search engine optimization professionals with previously unavailable information on user queries directed at Google’s generative AI features.

Anastasia Kourou first observed the phenomenon, reporting unusual queries such as ‘Yes’ and ‘Yes go on’ within her Search Console data. John Mueller, a Search Advocate at Google, later confirmed these were interactions with AI systems, according to a report by Search Engine Roundtable.

Suganthan Mohanadasan, writing for Search Engine Journal, detailed that a follow-up question within AI Mode is recorded as a new query. Any content generated in response to such a query, including impressions for linked pages, is attributed to it.

Google launched new Generative AI performance reports in June, which became universally available in August. These reports indicate AI visibility but do not include query data or clicks, and lack API support, Mohanadasan noted.

The author identified seven distinct types of AI conversation fragments within Search Console data. These included categories such as Reply Artifacts, Pivot Follow-Ups, Conversational Questions, Tracker Probes, Agent Harnesses, and Pasted Strings.

Links embedded within an AI Overview or AI Mode citation inherit the position of the entire AI block. This explains why certain single-word queries might show unusual average positions in performance reports, according to the analysis.

To assist users in managing this new data, a classifier has been integrated into a free Search Console MCP tool. This tool is designed to help identify and sort these AI conversation fragments, as reported by Mohanadasan.

The analysis of one site’s data over 16 months revealed 1,127 AI-related queries and 20,300 impressions. This indicates a small but measurable flow of AI interaction data into general Search Console reports, according to the findings.

Ross Tavendale, mentioned in the context of the analysis, contributed to understanding the implications of these findings for SEO strategies.

Source: Search Engine Journal