The running shoe question is personal and now conversational: bad knees, high arches, first marathon, described to an assistant with a budget attached. The answer names models, not just brands. Here is how those models get picked, a case study in how AI chooses products generally.
Shoe moments are body stories: the knee that complains past mile three, the plantar fasciitis comeback, the first race with a finish-time dream, the trail habit that outgrew road shoes, the replacement after five hundred miles. Runners state body, goal and budget, and assistants translate that into checkable criteria, stack height, support type, durability, price.
For shoe recommendations, assistants lean heavily on the category's unusually rich testing ecosystem, running publications that measure shoes in labs and miles, aggregated runner reviews, community threads where injuries and models are discussed together, and brand spec pages. Expert testing carries more weight here than in most consumer categories because it exists in depth and gets cited constantly.
Models that keep getting named are testable claims wrapped in corroboration: the stability shoe the tests call stable, the cushion the reviews call durable, the fit the threads call wide-friendly. Smaller brands take specific-need moments from the giants regularly, because a niche shoe's evidence matches a niche body question precisely.
The test costs nothing: ask ChatGPT, Gemini, Claude and Perplexity the questions above, phrased the way a real person would say them, and write down who gets named and what sources appear. Run each question twice on different days, since answers vary and patterns matter more than single runs.
For running brands, the fixes: spec transparency, presence in the testing publications assistants cite, and attention to the injury-adjacent threads where model reputations form. For running shoes that want the systematic version, the free Aethon audit runs your real buyer questions across all four assistants with screenshots, and the retail playbook turns the gaps into a work plan.
By converting body, goal and budget into criteria and matching them against expert testing, aggregated runner reviews, community threads and spec data. The category's deep testing ecosystem gives assistants unusually strong evidence to cite.
No. Shopping ads are separate and labeled. The models named in answers earned it through tests, reviews and community reputation.
Specific-need questions, wide feet, plantar fasciitis, ultra distances, match the niche shoe's test results and community reputation exactly, and assistants score the match.
Transparent specs, presence in the lab-testing publications assistants cite, and monitoring of injury and fit threads where model reputations form. The free Aethon audit shows which runner questions your models currently win.