Short answer: buyers ask AI assistants category questions, not brand questions. Before a procurement manager ever types your company name, they have asked tools such as ChatGPT, Perplexity, Claude and Gemini which companies make a certain kind of robot, how two suppliers compare, what a system roughly costs, and who can deliver and service it in their country. The answers to those questions — assembled from directories, company profiles, trade coverage and review pages — quietly become the shortlist. If your company is not in the underlying sources, it is not in the answer, and it never reaches the RFQ stage.
The five question families
- Selection: "Which companies make vision-guided bin picking systems?" "Who are the main makers of harmonic reducers?" These build the candidate pool, and they reward companies whose category membership is stated clearly and consistently in public sources.
- Comparison: "Supplier A vs supplier B for warehouse automation", "alternatives to [well-known brand]". Comparison answers are stitched from profiles and articles that state facts side by side — specifications, locations, product lines.
- Price: "How much does a robotic palletising cell cost?" Buyers rarely get a precise number, but they do get ranges and cost structures, and the suppliers named alongside credible pricing content gain a head start.
- Compliance: "Which robot suppliers can deliver CE-certified systems to the EU?" Regulatory facts — certifications, machinery-regulation readiness, safety standards — are exactly the kind of verifiable detail AI answers repeat.
- Service: "Who services collaborative robots in Germany?" Local presence, distributors and service networks decide these answers, and they are the questions closest to a purchase.
How the AI assembles its list
An AI assistant answering a selection question is not consulting a single database. It blends what it can find: structured directories, company profiles with concrete facts, trade-press coverage, and pages that answer the question directly. Two properties decide whether a company surfaces. The first is factual density — profiles that state category, headquarters, product lines and markets in plain language are easy to cite; pages of adjectives are not. The second is consistency — when your website, your directory profiles and your trade coverage describe the company the same way, the AI can repeat the description with confidence. Where sources disagree or say nothing specific, the answer fills the space with better-documented competitors.
The brand-name trap
Many suppliers test their AI visibility by asking an assistant about their own company — and get a fluent, reassuring answer, because the question itself supplies the name. That test measures nothing. The shortlist was already written by the category questions asked weeks earlier, when the buyer did not yet know your name. Visibility work therefore starts from the buyer's side of the table: the questions in the five families above, asked exactly as a buyer would ask them, with no brand prompt.
How to check where your company stands
The method our published guide to AI visibility for B2B suppliers sets out is deliberately unglamorous: fix a standard set of twelve category questions covering selection, comparison, price, compliance and service; run them unchanged across the main AI assistants; and record, per question, whether your company is mentioned, whether it is cited as a source, and which competitors appear instead. Repeat quarterly. The output is a map of exactly which question families you are missing from — which tells you whether the fix is a better profile, missing compliance facts, or absent trade coverage. Our step-by-step checklist for getting recommended by AI search turns the findings into an ordered work list, and the AI Growth service exists for suppliers who would rather have the audit run for them.
The uncomfortable summary: in AI answers, the shortlist is a citation contest, and it is won with specific, consistent, verifiable facts published where assistants look. Brand campaigns do not enter into it.
Frequently asked questions
What do buyers actually ask AI tools when researching robot suppliers?
Category questions, in five families: selection (who makes this kind of robot), comparison (how suppliers differ, alternatives to a known brand), price (what a system costs and why), compliance (who can deliver certified systems to their market), and service (who installs and supports robots in their country). Brand-name questions come later, after the shortlist exists.
Why does my company not appear in AI answers about its own category?
Usually because the public sources assistants draw on describe the company too vaguely or inconsistently. AI answers are assembled from directories, profiles, trade coverage and factual pages; a company whose category, products, location and certifications are not stated plainly and consistently across those sources gives the assistant nothing solid to repeat, so better-documented competitors fill the answer.
How can a supplier check its AI visibility properly?
Fix a standard set of category questions — selection, comparison, price, compliance and service — and run them unchanged across the main AI assistants, recording mentions and citations per question. Asking an assistant about your own company by name is not a valid test, because the question supplies the answer. Repeat the same question set quarterly to measure movement.
Does advertising or brand marketing improve AI shortlist visibility?
Not directly. AI shortlists are assembled from citable factual sources — structured profiles, directories, trade coverage and pages that answer buyer questions with specifics. Marketing spend may raise human awareness, but the citation contest is won by factual density and consistency: the same company facts, stated plainly, in every source an assistant might consult.