Research · preliminary

What 780 AI shopping answers show

From 1 to 5 October 2026 we asked 30 real US shopping questions about headphones and robot vacuums on ChatGPT and Google AI Mode, three times a day, and recorded every product recommended, its position and its price. These are the first results.

63% to 78%

of same-day repeat questions on Google AI Mode got a different first product recommendation (headphones and robot vacuums).

95% ranges: 47% to 76% and 59% to 94%
6% to 19%

of the time, ChatGPT and Google AI Mode recommended the same first product for the same question on the same day.

Robot vacuums 6% (1% to 14%), headphones 19% (7% to 35%)
18%

of Google AI Mode answers about robot vacuums showed a leading brand’s discontinued models, which shoppers can no longer buy new. ChatGPT: 1%.

95% range: 8% to 30%
1 in 7

robot vacuum prices on ChatGPT product cards did not match the retailer page they linked to, checked within hours of the answer. Headphones: 1 in 25.

Matched: 85% (82% to 89%) and 96% (94% to 99%)

What it means for a brand

  • One answer is not a measurement. The first pick changes between runs of the same question on the same day, so a screenshot says little about how often you win.
  • Each assistant needs its own measurement. ChatGPT and Google AI Mode rarely agree on the first product, so doing well on one says little about the other.
  • Catalog hygiene shows up in answers. Discontinued models and prices that don’t match the retailer page are recommended to shoppers today.

How we measured

  • Every run is a fresh, signed-out session in the US, in English
  • Questions were checked against Google search suggestions and Keyword Planner volumes, and the findings hold when questions are weighted by search volume
  • Prices on product cards were compared with the linked retailer page within hours
  • Ranges are 95% uncertainty ranges that treat repeated runs of the same question as related

Full methodology

Limits

Five days, two categories, one market and two assistants. Whether results drift over weeks is not yet measured. Brands are not named. The full study, with more questions and a longer window, will be published in November.

Get the full study

We will email you when it is published.

See your own products

A free audit applies the same measurement to your brand: where your products appear, how often they are the first pick, which prices are wrong and who takes their place. See a sample audit.