Why Most B2B Websites Are Invisible to AI Search (and How to Fix It)

IWD checked 99 B2B websites to see what ChatGPT, Google AI Overviews and Perplexity can actually reach and read. Most aren't blocking AI on purpose. They're losing it by accident.

Why Most B2B Websites Are Invisible to AI Search (and How to Fix It)

When a buyer asks ChatGPT, Google or Perplexity who supplies a part, the answer names a handful of companies. Whether yours is one of them depends on things most B2B teams have never checked. IWD Agency audited 99 B2B websites to find out what's going wrong, and turned the results into a full guide on how B2B brands get recommended by AI. This post covers the headline findings.

It matters because buyers have already moved. Gartner found that 45% of B2B buyers used generative AI during a recent purchase, mainly to research vendors and products. If AI doesn't name you while they research, you may never make the shortlist.

What IWD found in 99 B2B websites

IWD checked 99 public B2B websites across 10 sectors in October 2026, from the outside only, the way an AI crawler or a new buyer sees them. The short version: almost nobody blocks AI on purpose. The losses are accidental.

  • 1 in 8 turned AI away by accident. 12 of 99 sites let a normal browser in but turned away a request naming an AI bot, usually at the firewall or CDN. Only 2 of 89 readable robots.txt files block AI bots on purpose.
  • 1 in 4 hid the price in code. On 23 of 91 product pages that showed a price, the price only appeared after the page's scripts ran. Many AI crawlers read the page before scripts run, so they never see it.
  • 1 in 3 had no product data. 24 of 68 sites with measurable product pages had no Product structured data on the pages IWD checked.
  • Buying details were almost never stated. Minimum orders, payment terms and lead times appeared on 5 or fewer of 99 sites.

Why B2B catalogs get skipped

AI search tools do four things before they name a supplier: they reach your pages, read them, choose sources, and name a company. B2B catalogs tend to fail early in that chain.

The bots can't get in. Every major AI platform now separates its search bots from its training bots. Blocking a training bot like GPTBot doesn't remove you from ChatGPT answers, but blocking OAI-SearchBot, PerplexityBot or Claude-SearchBot does. Firewall and bot-protection settings often block them without anyone deciding to.

The page is unreadable. Part numbers, prices and specs loaded by scripts, or locked in PDFs, are invisible to crawlers that don't run JavaScript. OpenAI, Perplexity and Anthropic don't document running it, so the safe assumption is that they don't.

The products aren't online. Products sold through reps, dealers or quotes often have no page at all. When a buyer searches the exact part number, AI answers with a distributor's page instead of yours.

Nobody else mentions you. Industry research from Ahrefs found brand mentions on other websites track AI visibility about three times as closely as backlinks do. Distributor listings, directories and trade press carry real weight.

Five fixes you can start this week

  1. Check your robots.txt and firewall for OAI-SearchBot, PerplexityBot, Claude-SearchBot, Googlebot and Bingbot. Allow the search bots, and decide on training bots separately.
  2. View the source of a product page and search for the part number and price. If they aren't there, fix the product template, not every page.
  3. Show a public list price with "Sign in for your price" beside it if you have contract pricing.
  4. Publish one "How to buy" page that states who can buy, how to open an account, minimums, terms, shipping regions and lead times.
  5. Ask ChatGPT and Perplexity your buyers' real questions a few times each, and note whether you, a distributor or a competitor gets named.

What doesn't work

Some popular advice doesn't hold up. No major AI platform says it reads llms.txt, and Google says Search ignores it. A matched study of 1,885 pages found that adding schema markup produced no meaningful lift in AI citations. Structured data is still worth having for accuracy and shopping eligibility, but it isn't a shortcut.

Get the full method

The complete guide covers all seven steps, from crawler access to measuring what AI says about you, with a 14-point scorecard and a 90-day plan. Read how B2B brands get recommended by AI, or book a 30-minute assessment and the IWD team will run the checks on your catalog. For consumer stores, see our guide to AI search optimization for ecommerce.