If your company has a Facebook Page, an ad account or a login to Meta Business Suite, you are about to agree to something new. Meta's revised Commercial Terms take effect on October 30, 2026, and they include a paragraph that wasn't in the March 2024 version: permission for Meta's crawlers to visit, index and train on your website.
This is not a story about social media. It's a story about who gets to read your site, and what they do with it once they have. Meta runs an AI assistant that answers questions and a shopping agent that picks merchants. Both are built on what its crawlers bring back.
What the new clause says
According to reporting by PPC Land and Nieman Lab, a new paragraph in Section 1 has businesses authorize Meta, or anyone acting for it, to use "web scrapers, bots, robots, spiders, crawlers and user agents."
The scope is wide. It covers domains, pages and URLs "associated with your products or services." It allows Meta to access, index, cache, collect and crawl them, to pull updates periodically, and to use what it collects to train and develop its products. It also lets Meta create "test credentials" to get in, a term the document doesn't define.
What the clause doesn't say matters too. It doesn't define "associated with," doesn't mention robots.txt, and doesn't limit itself to product pages.
How you "agree"
Under Section 7, continuing to use Meta's business products after notice counts as acceptance. Nieman Lab reported there is no opt-out before October 30. Rejecting the terms means leaving Meta's business tools.
For most B2B companies that's not a real choice. The Facebook Page, the retargeting pixel and the ad account all sit under the same terms.
Meta says nothing has changed
Meta spokesperson Andy Stone called reports that the terms allow scraping news sites "false." He said the update concerns product information on third-party sites, used for ad and content recommendations, and that there is "no change" to how Meta uses crawlers. Content owners can still use robots.txt to state their preferences, he said.
A second spokesperson told Nieman Lab the crawling language was previously scattered across several documents and has now been consolidated in one place.
Lawyers read it differently
The legal experts Nieman Lab spoke to were less reassured. Matt Topic of Loevy & Loevy said it's unclear whether a court would enforce the clause, but that publishers "should assume that it would be enforceable." Cornell's James Grimmelmann read it as permission to train AI on publisher content.
The bigger worry is robots.txt itself. Compliance with it has always been voluntary. A signed contract is not. The experts said the new language could let Meta argue that its terms override a robots.txt block. Jason Kint of Digital Content Next put it bluntly on X: a robots.txt or IETF signal "shouldn't comfort any website."
Why this is about AI answers, not just training
It helps to know which Meta bots actually visit your site. Meta-ExternalAgent collects public pages for both model training and indexing, under a single robots.txt token. That means you can't block training without also blocking indexing. Meta-WebIndexer builds the index behind Meta AI's answers. FacebookExternalHit only fetches a page when someone shares a link, to build the preview.
That mix is the real problem for anyone who cares about AI visibility. Block Meta-ExternalAgent and you may drop out of what Meta AI knows about you. Allow it and your pages go into training too. There's no way to say "cite me, but don't train on me."
And Meta's crawling is no longer a rounding error. Cloudflare data cited by PPC Land shows Meta's share of AI crawler traffic rose from 0.9% to 7.5% between July 2024 and July 2025. DataDome found that Meta's Muse accounted for 33% of agentic browser traffic in September 2026. Muse opened to small businesses on September 29, and product browsing inside existing ads switches on October 28, two days before the new terms take effect.
The bigger picture
Every AI engine now has two decisions to make about your company: whether it can read your site, and whether it picks you when a buyer asks. Meta has just tied the first decision to your ad account. You no longer choose it page by page in robots.txt.
For most B2B companies, the right move isn't to block everything. It's to know which bots reach your site today, which ones feed answers and which only train, and what each engine actually says about you when a buyer asks. If Meta's crawlers are going to read your site anyway, the pages they find should be the ones you want quoted: clear product pages, honest comparisons, and answers to the questions buyers really ask.
Before October 30, check your server logs for Meta-ExternalAgent and Meta-WebIndexer. Then read your robots.txt as if a contract might outrank it, because now one might.
ClawWorld runs AI visibility work for B2B startups. Start with a free baseline report: 10 buyer questions, who ChatGPT and Claude name instead of you, and the pages they cite.