Final week, Search Engine Land printed knowledge from January to June 2025 associated to shopper conversion metrics in Google Search. The info implies that buyers are adapting their search habits to Google’s AI Mode: queries have gotten longer, conversions have universally decreased, and longer queries are capturing a bigger share of conversions. From the information:

• The Search share of queries of 1-2 phrases decreased from 42% in January to 31% in June, whereas queries of 3-4 phrases elevated meaningfully;
• Whereas click-through charges are down throughout the board over the timeframe, 1-2 phrase queries noticed a 50% lower in quantity, whereas queries of greater than eight phrases decreased by 26%.

As background, Google launched AI Overviews, that are LLM-generated summaries associated to Search queries, surfaced above the fold in Search outcomes, in Might 2024. Google adopted AI Overviews with AI Mode, its native Search chatbot, in Might 2025 (AI Mode was initially launched solely in the US and expanded to India this month). I’ve characterised Google’s implementation of AI-enabled merchandise in Search in my Google’s Gambit sequence (see half one and half two) as a change of the Search product expertise from a “Distribution Mechanism” to an “Engagement Sink.”

By this, I imply that the core performance of Google Search is evolving. Quite than rating hyperlinks to third-party web sites based mostly on the relevance of their content material to Search queries, Google goals to convey that info on to its customers. I articulated my interpretation of the motivation within the first submit within the Google’s Gambit sequence:

As a distribution mechanism, Google optimizes to generate one click on per search: a consumer considers a search question profitable if it results in a click on to the suitable vacation spot from the primary web page of outcomes. However the incentive with AI Overviews may be very clearly completely different: it’s to resolve the consumer’s wants with out resulting in a click on, doubtlessly whereas instigating additional queries and, subsequently, elevated time spent with AI Overviews. And advertisements are displayed every time AI Overviews are rendered, doubtlessly main to 1 click on per question because the consumer refines their search. In impact, with AI Overviews, Google can seize the conversions and advert clicks that happen subsequent to a profitable search question within the present distribution mechanism configuration.

Over the weekend, I shared knowledge from a survey performed by Oppenheimer & Co. on X. I used to be requested to not share your complete report, however I discovered the 2 graphs above illuminating: of customers aware of AI Mode and who pay for ChatGPT, 82% discover AI Mode extra useful than Google Search and 75% discover AI Mode extra useful than ChatGPT. This means that AI Mode will not be solely fulfilling the objective I hypothesized in Google’s Gambit, however it’s additionally staving off the aggressive risk from chatbots to Search. Moreover, AI Mode seems to be encouraging customers to make the most of extra pure language with queries in Search, with the lower in click-through charges presumably coinciding with extra time spent on web site, and a correspondingly increased variety of advertisements served.

To this finish, Google’s Gambit seems to be working, with the principal practical use case of Search transitioning from discovery to content material engagement. Not solely that: customers want this new format to conventional Search. The implications of this for the open net are ominous: as I write in Google was the open net’s imperfect benefactor, Google has little motive to assist the open net whether it is compelled to promote its Community enterprise, which it could be because of its “advert tech” antitrust trial.

However even when not, Google should forestall the verb “Google” from turning into an anachronism as customers flip to chatbots for basic info gathering. AI Mode is the pure technique to make sure that Google’s Search enterprise stays the web’s principal level of knowledge discovery. This places scaled net publishers in a tough place: in the event that they want to retain the inbound Google site visitors on which their companies have been constructed and basically rely, they could have to fulfill Google on its new phrases: to make their knowledge obtainable for coaching in trade for visibility in AI Overviews.