Can AI see your B2B catalog?

Buyers now start supplier research inside AI assistants, and the answer they get becomes their shortlist. Most large B2B catalogs aren’t in it. This field guide, written with Prerender.io, shows the two walls keeping them out, how to tell which one you’re hitting, and a six-step plan to clear both without rebuilding your platform.

The Invisible Catalog

21 pages, PDF

Get the field guide

The Invisible Catalog · 21 pages, PDF

A ten-minute self-test, a readiness checklist and a six-step plan.

Business email^*

Company^*

I consent to receive future marketing communication from Virto Commerce group.

I confirm that I have read & agreed to Terms & Privacy Policy. ^*

Two fields. Nothing else is required.

What changed in the last 12 months

AI crawlers arrived fast, and most catalog pages aren’t built for them. Per Prerender.io’s crawl data:

AI crawler traffic grew about 3.7x in a year across 20 enterprise retailers

AI crawlers now make up a quarter to a third of all AI and search crawl activity, up from about 10%

93% of 3 million sampled catalog pages took longer than 5 seconds to fully load

69% took longer than 10 seconds, which is too slow for a crawler on a limited budget

GPTBot, PerplexityBot, ClaudeBot and similar crawlers read the raw HTML your server sends and don’t run JavaScript. On a JavaScript-heavy catalog, they get a page with no products, no prices and no specs.

In B2B, a product only exists in the context of a buyer. So the question is never just ‘can AI see my catalog.’ It’s ‘which version of my catalog should AI see, and where does that stop.’

Oleg Zhuk

CTO, Virto Commerce

Written for everyone who owns a piece of the problem

Digital, SEO, IT and catalog owners each hold part of the answer. The guide gives them the same diagnosis and the same plan to work from.

Two walls, explained plainly

Why AI crawlers can’t read many catalog pages, why they can’t reach the data that matters most in B2B, and why most teams only ever work on the first.

A ten-minute self-test

Ten buying-intent queries, three quick technical checks and one entity check. Together they show whether AI names you, skips you or describes you wrong.

Public layer, private layer

How to decide, attribute by attribute, what AI should see and what stays behind the login. Buyers and crawlers get the same public page, and contract pricing stays with the accounts entitled to it.