Your website just got a new choice, and it is one most owners have never been offered before: who is allowed to learn from your site, and who is allowed to quote it to a customer. Those are two different doors, and this week's Scale with AI episode explains why the answers should probably be different, what the new shared opt-out signal actually does, and where its limits are. Listen to the episode: Scale with AI — October 31, 2026.
The Shift
Four of the biggest AI companies agreed this week to honor a single, shared signal that lets a website owner say, in one place, do not use my content to train your models. Every website carries a small instruction file for automated visitors, a sign on the door that robots read before people do. Until now, an owner who cared had to chase every crawler separately. The new agreement adds one common line that all four say they will respect.
But training your content into a model and quoting your content to a customer are two completely different things. Training means your words help build the system itself. Quoting means a customer asks who to call in their town, and the assistant fetches your page in that moment and recommends you by name, with your hours, your services, and your phone number. Block everything and you have protected your content and disappeared from the answers. Block nothing and you gave away the training value. The smart position is the middle one: decline training if that is your choice, and say yes, loudly, to being found, quoted, and recommended, with a site built for it: answer-first pages, clear facts, consistent names and numbers.
The Law
Owners deserve the limits in plain English. This agreement is a voluntary pledge, not a law; nobody fines a company for ignoring it today. It arrives on a staggered schedule into 2027, covers only future training, does not undo anything already collected, and does not stop a visit to your site for search or for answering a question. The companies themselves warn that turning off training may also lower the chances your content shows up inside AI experiences, which is why this should be a deliberate decision made with advice, not a switch flipped because a headline scared you. The calling rules are unchanged: consent, honest identification, an immediate opt-out, and legal calling hours, and in Washington everyone on a call must agree before it is recorded.
Who's Doing It
The story this week is the gap itself. A major October survey of more than two thousand small businesses found the overwhelming majority now use AI somewhere, and almost all say it helps, yet only about one in six of those using it say AI is woven into their core operations. The gap shows up two ways. First, the invisible owner: a competitor keeps getting recommended by assistants because their site was rebuilt answer-first with clean, consistent facts and deliberate training and discovery settings, while yours was built only for human eyes. Same company, same reviews, different front door. Second, shadow AI: staff quietly adopt tools nobody approved, with customer information flowing into accounts nobody vetted. The businesses getting this right chose a small number of tools on purpose, set the rules in writing, and integrated them into how the work actually flows.
Your Move
Three steps. First, decide your website's position on purpose: training, yes or no; being quoted and recommended, yes or no. Both answers should be yours. Second, make your site quotable: the answer first on every important page, and your name, services, service area, and phone number identical everywhere they appear, because an assistant can only recommend you confidently when your own facts agree. Third, inventory the AI tools already inside your business, including the unofficial ones, and approve or retire each in writing. If you want that built for you, that is exactly what we do at The Agency. Start with a free AI Audit and we will show you what an assistant sees when it looks at your business today.