Cloudflare starts blocking agents by default as Google tests paying publishers
Cloudflare's September deadline landed today, and it decides who gets to read a lot of sites. Google started paying a handful of publishers when AI answers use their content, and gave every business a tab showing what its automated assistant has written down about them.
This Week's Signals
Cloudflare blocks training and agent bots by default starting today
Why it matters for your score
Cloudflare said in July that from 15 September, training and agent bots would be blocked by default on pages that display ads, while search bots stay allowed. It applies to domains newly onboarding rather than established paid customers. If your site landed on Cloudflare recently, agents may now be shut out of your ad-serving pages without anyone choosing that, and the fix is a setting rather than a rewrite.
Google starts paying some publishers when AI answers use their content
Why it matters for your score
Google confirmed an early-stage AI contribution pilot that pays sites whose content contributes to Gemini, AI Overviews and AI Mode answers, with earnings shown in a Search Console widget. Publishers call the payout formula black box and the early amounts small. Being cited is starting to carry a number next to it, and the pilot suits small and mid-sized sites better than large ones.
Google Business Profiles get a Collected info tab you can edit
Why it matters for your score
Google now shows a Collected info section where you review details its automated assistant gathered by call, text or WhatsApp, see the source and date, and delete what is wrong. Those details feed what assistants say about your business, so an error you never approved is an error they will repeat.
Microsoft Advertising sets disclosure rules for AI generated ads
Why it matters for your score
Microsoft published guidance requiring advertisers to disclose AI created or materially altered ads and to preserve watermarks, metadata and provenance, with rejection for missing disclosures or prohibited deepfakes. The provenance data your tools attach is now load-bearing, so stripping it on export costs you the ad.
Google tests sponsored text links above AI Mode answers
Why it matters for your score
Google is testing plain text links marked Sponsored above the AI Mode response, styled to sit with the rest of the answer rather than in a separate ad block. Paid placement is moving into the part of the answer people actually read.
Emerging Audit Signals
Access is the question this week, not content.
Two of these 5 stories are about who is allowed to read your site at all.
Payment for AI use is arriving before measurement of it.
Google pays on a formula publishers cannot see.
Machines call your business now, they do not only crawl it, and what they write down becomes what assistants repeat.
Ad inventory keeps moving inward.
First AI Overviews, now the AI Mode answer itself.
What to do this week
- 1
Check your Cloudflare bot settings today. If your domain onboarded recently, confirm whether the Training and Agent categories are blocked on your ad-serving pages, and decide deliberately instead of inheriting a default.
- 2
Open your Google Business Profile, go to Edit profile then Collected info, and delete anything the automated assistant recorded wrong.
- 3
Take your 5 highest-intent pages and load them with JavaScript disabled to confirm the answer sits in the HTML an agent receives. Faro's /tools/ai-readiness-scan runs the same check across the site in one pass.
Sources
- ↗ Your site, your rules: new AI traffic options for all customers
- ↗ Cloudflare's new policy pushes AI companies to pay for publishers' content
- ↗ Google rolls out pay-per-use AI licensing program to publishers
- ↗ Manage information collected about your business, Google Business Profile Help
- ↗ Microsoft Advertising sets rules for AI-generated ads
- ↗ Google AI Mode Tests Text Link Ads