TLDR;
Google introduced Personal Intelligence in AI Mode for eligible US subscribers who choose to connect Gmail and Google Photos. Personal context can change the recommendations someone receives. Brands need clear product relevance and should avoid treating one personalised answer as the result every customer will see.
What happened
Google introduced Personal Intelligence in AI Mode for eligible US subscribers, using opted-in Gmail and Google Photos context to tailor answers. The experiment made recommendations more dependent on an individual’s circumstances, with explicit connection controls rather than assuming every search user participates.
Why it matters
Connected account information can affect what is suitable for an individual, introducing context that a brand cannot observe through a conventional rank check. The enterprise implication is stronger product relevance and more careful measurement. There is no sound basis for treating one personalised answer as representative of every eligible user.
An opt-in assistant using connected accounts can understand context that a public query alone may not reveal. That can change the recommendations a person receives, but it does not mean a brand gets access to the underlying Gmail or Photos information. Enterprises should resist turning a consumer personalisation feature into an assumed targeting opportunity. The more practical implication is that generic visibility tests may become less representative of individual customer experiences. Different context can produce different priorities even when the public wording of a request looks similar. That reinforces the value of explaining product suitability and limits clearly. A brand cannot control a person's private context, but it can make the public facts needed for a good match more complete and accurate.
How your brand can benefit / be affected
Make objective relevance signals explicit: geography, availability, compatibility and eligibility. Avoid inventing content around assumed private account data or implying the brand can target that information.
Keep generic and personalised observations separate, recording the context and market used in legitimate tests. Evaluate whether the answer is correct for the scenario. Use consented customer research to understand journeys that analytics cannot see, without trying to reconstruct private user context.
Review whether your content explains the circumstances that change suitability. A service may fit someone travelling frequently but be inappropriate for a customer with a different usage pattern; a product may solve one constraint while introducing another. State those distinctions plainly and avoid claiming that the offering suits everyone. Use existing customer research to identify relevant scenarios rather than inventing personal profiles based on speculation about connected data. Make the local market, eligibility and commercial conditions easy to identify. These improvements support contextual interpretation without requiring the brand to collect additional personal information or depend on knowledge it does not receive.
Treat testing as scenario research rather than a universal ranking report. Compare realistic questions and note the account, market and feature conditions under which you observed the response, using appropriate test data and permissions. Do not put real customer information into an experiment simply to simulate personalisation. Evaluate whether the source facts remain correct when the recommendation changes. Separate what you observed from what you infer about the wider audience. If reporting cannot quantify personalised exposure, say so. The enterprise benefit comes from serving the right use cases well and avoiding misleading claims, while recognising that an opt-in feature has a narrower scope than all Google Search users.
News date: 22 January 2026. Editorial review: 16 September 2026. Analysis includes subsequent developments where stated.