94% of B2B buyers find suppliers with AI tools – how visible are the TIK participants there?
We tested 62 websites of TIK participants. How visible are they in ChatGPT, Gemini, Perplexity and Microsoft Copilot? We fired off all 12,250 prompts and analyzed the answers.
How visible the TIK companies are in the major LLMs:
- 62
- B2B industrial companies, each its own category
- 12,250
- AI answers analyzed
- 116,062
- deduplicated brand mentions
- 118,099
- citations
If you are not named, you are not asked.
According to a Forrester study, the share of B2B decision-makers using AI systems for search has risen to 94%. If a company does not appear in the answer, its chance of an inquiry drops.
Inquiries often consider only 3–5 companies. Whoever does not make it into the AI answer does not exist for the buyer in that moment.
Even the most visible brand holds only a small share.
Share of Voice is a brand’s share of all brand mentions in its category. The highest value is 17.2%. Not a single participant reaches 20%.
Distribution across all 62 participants
On average, participants hold 6.6% of their own category. Only 11 of 62 reach 10% at all, and none exceed 20%.
Highest self share of voice per participant
- BGS Beta-Gamma-Service17.2%
- Brückner Maschinenbau15.0%
- Julius Blum13.4%
- Liebherr-Rostock11.4%
- ebm-papst11.1%
- E-T-A Elektrotechnische Apparate11.1%
- Trumpf10.8%
- Komax10.7%
- Sennebogen10.7%
- Cognex10.1%
- Würth9.8%
Even on home turf, a competitor usually leads.
Every question targeted the participant’s core business. Still, in 40 of 62 categories the most-named brand is not them but a competitor.
participants total
in the top 10 of their category
in the top 3
the most-named brand
27 of 62 do not lead in a single one of the four models. 13 do not even rank in the top 10 of their own field, and 2 appear in no answer at all.
Your brand can be visible and still not win. What counts is your share of the answer.
More detail on your visibility – test the tool now
ucited measures your visibility across ChatGPT, Gemini, Perplexity and Copilot — including competitors, citations and concrete gaps.
Brand share in AI answers, not Google rankings.
Ten neutral category questions per participant, sent verbatim and without brand names. Each question targeted the participant’s core business. Competitors were discovered from the answers, not predefined.
- 62
- participants, each its own category
- 10
- neutral prompts per participant
- 4
- AI models: ChatGPT, Gemini, Perplexity, Copilot
- 5
- iterations per prompt and model
Ask the same question five times — and the answer changes.
We sent each question to the same model five times. Even brands that appear at all are missing from all five runs in 46% of cases. And the more a model searches the web live, the shakier it gets.
Share of unstable mentions per model
- Copilot70.6%
- Perplexity41.6%
- ChatGPT38.0%
- Gemini24.9%
Gemini did not search live in this dataset and is the most stable. Copilot searches live almost always — and most often makes brands disappear between two identical runs.
That is why a one-off check tells you almost nothing. Visibility in AI answers has to be measured continuously.
Each model tells a different story.
Visibility in ChatGPT does not automatically mean visibility in Gemini, Perplexity or Copilot. And more live search does not mean more brand visibility.
- Avg brands / answer
- 10.17
- Participants invisible
- 3
- Avg brands / answer
- 7.47
- Participants invisible
- 11
- Avg brands / answer
- 9.27
- Participants invisible
- 5
- Avg brands / answer
- 10.96
- Participants invisible
- 4
Perplexity triggers a live web search almost always (99.7%), yet names the fewest brands per answer and makes 11 participants disappear entirely in their own category. Gemini never searched live in this dataset (0%) but names the most brands. Live search and visibility are not the same thing.
Across all brands, 84.5% are missing from at least one of the four models, 69.1% from at least two. Only 15.5% appear in all four. Of the 62 participants, 14 are invisible in at least one model, and 2 in all four.
The AI has incumbents — and they bleed into other categories.
A few large brands appear across many categories. Specialists therefore compete not only with their direct rivals but with the corporations the model knows as a safe answer.
- Siemens1,640in 28 categories
- Bosch803in 16 categories
- ABB803in 17 categories
- Schneider Electric639in 15 categories
- Beckhoff512in 6 categories
- Emerson495in 11 categories
- KSB452in 5 categories
- Liebherr436in 6 categories
- Festo431in 8 categories
- Phoenix Contact410in 11 categories
Siemens alone is named 1,640 times, spread across 28 of the 62 categories, and is often the category leader of other participants.
Whoever is not present in their category cedes it to the brands the model knows as a safe answer.
The answer is built from third-party sources, not your website.
B2B AI answers lean on structured external sources: supplier portals, industry directories, trade sites, Wikipedia and Reddit.
Most-cited domains
- 1wlw.de5,487
- 2reddit.com3,115
- 3chemie.de2,782
- 4de.wikipedia.org2,666
- 5induux.de1,882
- 6mordorintelligence.com1,696
- 7europages.de1,595
- 8en.wikipedia.org1,266
The ten most frequent domains together account for 18.8% of all citations. Wikipedia reaches 3,932, Reddit 3,115, LinkedIn only 270.
Six source types carry the answers
Only 5% of citations point to the company’s own website.
Of 118,099 citations, only 5,955 point to the participant’s own domain. Onpage is the base — offpage is what the AI explains your brand from. Citation gaps are the new content gaps.
From finding to fix.
GEO is not a one-off audit. Five levers to measurably improve your position in AI answers.
- 01
Check citation coverage
Which top sources cite your competitors but not you? Is your brand missing from wlw, induux, europages, industrystock or relevant trade portals — and are the entries current and semantically correct?
- 02
Strengthen semantic anchoring
Is your brand tied to the right attributes — product categories, certifications, industries, applications, locations? And does it appear next to the right competitors in comparison contexts?
- 03
Build an entity hub on your own site
A clear central brand page, Organization/Brand schema, clean product and category structure, crawlable and renderable content. Technical SEO stays the foundation — not the whole story.
- 04
Improve partner, distributor & directory texts
In B2B especially, distributor, reseller, association and marketplace profiles provide the external confirmation AI systems use to place a brand. These texts are GEO-relevant.
- 05
Measure repeatedly
GEO is not a one-off audit. Measure a baseline, close gaps, improve external profiles, measure again — and track the change in share of voice.
What share does your brand hold in AI answers?
See your own numbers from day one — visibility, competitors, citations and concrete gaps across all four models.
ucited.ai is morefire’s GEO tool. Come talk to us at the bvik TIK 2026.
- Market and language: Germany, German, B2B industry. The findings apply to this context and do not generalize to other markets or sectors.
- Share of Voice is a brand’s share of all deduplicated brand mentions in its category. Each brand is counted once per answer.
- Gemini triggered no live web search in this dataset (0%) and answered from model knowledge. This is not a data error.
- As of June 2026. All figures were recomputed directly against the raw analysis database.