What technical buyers actually ask AI before they shortlist a supplier

A procurement engineer needs a supplier for a part, a material, a sub-assembly. Ten years ago they’d have asked a colleague or opened Google. Today, more and more of them open ChatGPT, Gemini or Perplexity and type something like: “Who can machine 5-axis titanium components to aerospace tolerances in the UK?”

The assistant answers with three or four names. Those names are the shortlist. Everyone else was cut before a human looked at a single website.

That’s the part most manufacturers haven’t clocked yet. AI visibility isn’t a marketing gimmick bolted onto SEO — it’s a new first-round filter on your pipeline, and it runs whether or not you know it exists.

The question behind the question

When a buyer asks an assistant to recommend a supplier, the model isn’t reasoning about your engineering. It’s assembling an answer from what it can read and trust about you across the web. So the real question isn’t “is my site nicely designed” — it’s “can a machine that has never met me state, in one sentence, exactly what I make and for whom?”

Here’s what that machine is effectively checking:

• Can it tell what you actually do? Not “precision engineering solutions” — the specific processes, materials, tolerances, sectors. Vague brochure language gives the model nothing to match a query against.
• Can it place you? Location, the markets you serve, whether you take the kind of work being asked about.
• Is there corroboration? A single self-published claim is weak. Structured data, consistent descriptions, third-party mentions and clear capability pages give the model reasons to name you with confidence.
• Can it quote you? Assistants love content that answers a question cleanly — a capability spec, a straight FAQ answer, a plainly-stated standard. Prose that dances around the point doesn’t get cited.

If your website is a slideshow of factory photos and the phrase “your trusted partner,” you are invisible to this process. Not low-ranked. Invisible.

Why your best salesperson can’t help here

The cruel bit is that this filter runs before the stage where you’re good. Plenty of manufacturers win the job the moment they get a human on the phone — the depth shows, the trust builds, the quote lands. But the AI shortlist happens upstream of all that. If the assistant doesn’t surface you, the buyer never calls, and you never find out you were in the running. There’s no bounced enquiry to analyse. The loss is silent.

That’s what makes this different from ordinary “we should do some SEO.” A page-two Google ranking still lets a determined buyer scroll to you. An AI answer that omits you gives the buyer no reason to look further — they’ve got their three names.

What actually moves it

The fix isn’t clever prompt-stuffing or a chatbot on your homepage. It’s boringly fundamental, which is exactly why so few competitors have done it:

1. Say what you do in the buyer’s words. A page per process, material and sector, stated in the terms a buyer would actually type — not the terms your marketing agency preferred.
2. Make your claims machine-readable. Structured data (schema) so the specifics — what you make, where, to what standard — are legible to search engines and assistants, not buried in an image.
3. Answer the real questions plainly. The questions a buyer asks before enquiring, answered straight, in a way an assistant can lift and quote.
4. Measure it. You can’t improve what you can’t see. It’s genuinely possible now to check how often the assistants recommend you, for the exact questions your buyers ask, and who they name instead. (That measurement is the whole basis of how I approach this — you fix what the numbers show, not what you guess.)

None of that requires you to become a content machine or chase trends. It requires your website to state, clearly and consistently, the true thing about your business — so a machine sorting suppliers on a buyer’s behalf has something real to go on.

The manufacturers who get this right in the next couple of years won’t necessarily be the best engineers in their field. They’ll be the ones a buyer’s AI assistant can actually describe. Those are not always the same companies — and that gap is the opportunity.

If you want to know where you currently stand, that’s exactly the kind of thing worth measuring before you change anything. Book a 15-minute call and I’ll show you what the assistants say about you today.