Key points
- Google's Ads Decoded episode of October 28 featured Ginny Marvin and Shivani Mohan discussing how people use AI Mode and plan purchases.
- Google says AI Mode queries run about three times longer than traditional ones; its U.S. report puts multimodal use at over one in six AI Mode searches.
- Google has not published conversion benchmarks, industry breakdowns or buying-stage data for these behaviors, so intent claims remain untested.
- Mohan says users value AI summaries but still seek forums, podcasts and videos for lived experience.
Why it matters for marketers: Longer, exploratory questions mean ad copy, landing pages and reviews must answer real customer concerns earlier, but without Google benchmarks, teams must test for themselves what drives purchases.
Google used its October 28 Ads Decoded episode to describe how people search in AI Mode. Ginny Marvin, the company's Ads Product Liaison, spoke with Shivani Mohan, who leads data science and user experience research for Search. They said people ask longer questions, use voice and images, and start exploring decisions well before they buy.
Google cites a figure of 3x longer queries in AI Mode compared with traditional search. Its U.S. report also says multimodal searches make up more than one in six within AI Mode. Marvin's LinkedIn post phrased that more broadly, so the figure should not be applied to all Google searches. The longer-query number compares average length only and does not show stronger buying intent or better conversion.
Mohan said users like the clarity of AI summaries but still want detail, experience and authority from the open web, including forums, podcasts and videos. For marketers, that suggests product copy alone may not answer what buyers want to know: how something performs, where it falls short and why someone chose it.
The source material offers no campaign settings and no breakdown by industry or buying stage. Suggested checks include reviewing customer questions and search terms for needs described before people know a product name, making price and setup details easy to find on landing pages, and linking reviews or demos to known concerns. Teams should also track whether research visits lead to later purchases, rather than relying on impressions or click-through rate alone.
Sources
This brief is our own summary. Read the full reporting at the source.
