The moments that trigger the question
Contractor moments split between emergencies, the burst pipe, the storm damage, the failed furnace, and projects: the kitchen, the addition, the commercial build-out. Emergencies ask who can come now and be trusted; projects ask who does quality work at fair prices. Both carry fear of being cheated, which makes trust evidence the whole game.
The sources the answers lean on
For contractor recommendations, assistants verify license and insurance records, review platforms with volume and recency, local directories, and contractor sites that explain process and pricing honestly. Photos prove little to a model; a page explaining what a roof replacement involves and costs in your region is quotable evidence. Review specifics, showed up on time, cleaned up, fixed the callback, become the framing the assistant repeats.
Why the same contractors keep winning
The contractors that keep getting named are verifiable: licensed and displayed as such, consistently listed, reviewed recently and specifically, with content that answers what worried owners ask. Small operations beat big-brand franchises constantly in this category, because the evidence that matters, local reviews and license records, is native to small operators who tend it.
Run the test yourself, and what contractors can do about it
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 contractors, the fixes: display license and insurance in machine-readable form, reconcile every directory, publish honest process-and-cost pages per service, and treat review flow as a weekly habit. For contractors that want the systematic version, the free Aethon audit runs your real buyer questions across all four assistants with screenshots, and the construction playbook turns the gaps into a work plan.