Industrial sourcing questions moved into assistants quietly: an engineer describes the part, the material, the tolerance, the volume and the deadline, and asks who can make it. The answer names manufacturers. Here is what gets a shop onto that list.
Sourcing moments are project events: the prototype that needs ten units by Friday, the supplier that failed an audit and must be replaced, the reshoring mandate, the material switch, the capacity crunch before a launch. Buyers are specific, process, material, certification, geography, and assistants match against exactly those specifics.
For manufacturer recommendations, assistants read industrial directories and their structured capability data, certification records, ISO and industry-specific, trade press, and manufacturer sites that state capabilities in checkable terms: machines, materials, tolerances, capacities, certifications. Capability pages written as specification answers get quoted; we deliver excellence pages do not.
The shops that keep appearing publish their capabilities like a datasheet and are corroborated by directories and certifications that say the same thing. Small specialized shops beat large generalists whenever the ask carries specifics, because five-axis titanium with AS9100 is checkable and either matches or does not.
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 manufacturers, the fixes: capability pages in specification language, directory profiles that match them exactly, certifications displayed in machine-readable form, and case-shaped content for the industries you want. For manufacturers that want the systematic version, the free Aethon audit runs your real buyer questions across all four assistants with screenshots, and the manufacturing playbook turns the gaps into a work plan.
By matching stated specifications, process, material, tolerance, certification, geography, against checkable evidence: directory capability data, certification records, trade press and datasheet-style capability pages.
No. Directory ads buy directory placement only. The recommendation text is synthesized from verifiable sources on every major assistant.
Because the ask was specific and the small shop's published capabilities matched it exactly. Assistants score specification match, which specialists state more precisely than generalists.
Capability pages written like specifications, directory profiles that mirror them, and machine-readable certifications. Then measure the four assistants on your sourcing questions monthly. The free Aethon audit provides the baseline.