Cohort: Carpathia Top 50 B2C Onlineshops Schweiz 2026Scanned 26 Aug 2026Ranked by score, never by payment
Switzerland’s biggest shops are the least ready for AI agents.
All 50 shops on the 2026 Carpathia revenue ranking, scanned for the technical basics an AI agent needs to read and act on a site: llms.txt, sitemap, robots policy, structured data, and whether AI user-agents are let in at all. Nobody in the revenue top 10 scores above 60. 3 of the ten biggest shops in the country did not respond to the scan.
3
of the top 10 by revenue would not respond to the scanner at all
34.6
average score of the 7 top-10 shops that could be scanned
26%
of scannable shops refuse every AI user-agent probed
#15
revenue rank of the shop with the highest comparable score (85)
The board
Score out of 100 across 16 scored checks. Revenue rank is the shop’s position on the Carpathia list. Ranked shops come first: the 28 where every scored check produced a verdict. Below them, 18 with a score but at least one check we could not read — hatched, and deliberately unranked, because a score over fewer checks is a different measurement. At the bottom, 4 that did not respond at all: an absence of a result, not a low one.
| # | Shop | Rev. | Readiness | Score | Notes |
|---|---|---|---|---|---|
| 1 | apple.com | #15 | 85 | ||
| 2 | ikea.com | #12 | 85 | ||
| 3 | exlibris.ch | #21 | 83 | ||
| 4 | apfelkiste.ch | #26 | 79 | ||
| 5 | ifolor.ch | #37 | 79 | ||
| 6 | baby-walz.ch | #36 | 77 | ||
| 7 | nettoshop.ch | #21 | 77 | ||
| 8 | qoqa.ch | #16 | 75 | ||
| 9 | aboutyou.ch | #24 | 70 | ||
| 10 | landi.ch | #41 | 70 | ||
| 11 | ackermann.ch | #44 | 69 | ||
| 12 | fust.ch | #24 | 69 | refuses 1 of 4 agents | |
| 13 | ottos.ch | #21 | 69 | ||
| 14 | manor.ch | #27 | 67 | ||
| 15 | aliexpress.com | #11 | 62 | ||
| 16 | bestsecret.ch | #17 | 60 | ||
| 17 | bonprix.ch | #30 | 60 | ||
| 18 | globus.ch | #45 | 60 | ||
| 19 | nespresso.comtop 10 | #10 | 60 | ||
| 20 | pfister.ch | #49 | 60 | ||
| 21 | zooplus.ch | #30 | 60 | ||
| 22 | lehner-versand.ch | #38 | 50 | ||
| 23 | hornbach.ch | #27 | 45 | ||
| 24 | ochsnersport.ch | #34 | 44 | refuses 3 of 4 agents | |
| 25 | amazon.detop 10 | #5 | 40 | ||
| 26 | amazon.fr | #20 | 40 | ||
| 27 | amazon.it | #43 | 40 | ||
| 28 | zalando.chtop 10 | #2 | 30 | refuses 2 of 3 agents | |
| — | ch.shein.com | #13 | 86 | 1 of 16 unread | |
| — | conforama.ch | #42 | 83 | 1 of 16 unread | |
| — | ideoon.ch | #48 | 64 | 1 of 16 unread | |
| — | iqos.com | #40 | 64 | 1 of 16 unread | |
| — | laredoute.ch | #36 | 63 | refuses 1 of 4 agents · 2 of 16 unread | |
| — | temu.comtop 10 | #4 | 45 | 1 of 16 unread | |
| — | decathlon.ch | #30 | 25 | blocks agents · 2 of 16 unread | |
| — | interdiscount.ch | #14 | 25 | blocks agents · 2 of 16 unread | |
| — | mediamarkt.ch | #18 | 25 | blocks agents · 2 of 16 unread | |
| — | migros.chtop 10 | #9 | 25 | blocks agents · 2 of 16 unread | |
| — | notino.ch | #50 | 25 | blocks agents · 2 of 16 unread | |
| — | ricardo.chtop 10 | #6 | 25 | blocks agents · 1 of 16 unread | |
| — | conrad.ch | #33 | 17 | blocks agents · WAF · 4 of 16 unread | |
| — | coop.chtop 10 | #8 | 17 | blocks agents · WAF · 4 of 16 unread | |
| — | deindeal.ch | #35 | 17 | blocks agents · WAF · 4 of 16 unread | |
| — | douglas.ch | #46 | 17 | blocks agents · WAF · 4 of 16 unread | |
| — | jumbo.ch | #39 | 17 | blocks agents · WAF · 4 of 16 unread | |
| — | orellfuessli.ch | #29 | 17 | blocks agents · WAF · 4 of 16 unread | |
| — | galaxus.chtop 10 | #1 | — | no response | |
| — | digitec.chtop 10 | #3 | — | no response | |
| — | brack.chtop 10 | #7 | — | no response | |
| — | hm.com | #19 | — | no response |
What “blocks agents” means
12 shops answer 403 to every AI user-agent probed — GPTBot, ClaudeBot, Google-Extended and PerplexityBot alike. 6 of them refuse even robots.txt, so there is no policy to read: the wall answers before the file does. Their scores are a ceiling, not a measurement — most of what the checks look for could not be reached, so a shop in this group and a shop that is simply open and unprepared are not on the same scale.
A further 4 refuse some agents but not others. They carry the count rather than the badge: refusing three of four is a different decision from refusing all four, and collapsing them into one label is how a term stops meaning anything.
Blocking may be a deliberate decision. The board does not judge the decision, only reports it.
Method
- One scan per shop on 26 Aug 2026, from a single address, using the same checker anyone can run here. Results are cached; this page is a snapshot, not live.
- 16 scored checks across llms.txt, sitemap, robots and AI user-agent policy, JSON-LD and
.well-known. Forward-readiness checks are reported and never scored. - A shop is ranked only when every scored check produced a verdict. One unread check and it keeps its score but leaves the ranking — a score over 16 checks and a score over fewer are not the same measurement.
- Sites that answered nothing are listed without a score. We do not publish a zero for silence.
- Scores can be disputed line by line: every failed check names the URL and the status it returned.
- Nothing on this page was paid for and nothing can be.
Not on the list? Run the same 16 checks on your own shop. It takes about a minute and shows exactly which line failed.
Scan your shopCohort from Carpathia Digital Business Insights. Revenue ranks are theirs; readiness scores are ours.