Most vendor announcements describe what is possible. Microsoft’s October 30 announcement of separate AI training and AI grounding preferences also states a cost. It says that restricting grounding reduces brand visibility in AI experiences. That sentence will carry a lot of weight in internal meetings, so it is worth reading closely for what it supports and what it leaves open.
Start with what it supports. Microsoft is telling publishers that, in its view, content used as input to help inform generated responses is part of how a brand shows up in Microsoft’s AI experiences, and that withholding it has a visibility consequence. That means the decision is not free, and a team that treats restricting grounding as costless has not read the announcement.
Now what the statement leaves open. It does not give the size of the effect. It does not say which pages matter most, which kinds of questions are affected, or whether visibility in an AI experience turns into traffic, leads or sales. It is a vendor’s statement about its own product, and no measurement of it appears in the announcement or in this article.
The announcement is just as plain about the other side. Microsoft says neither choice guarantees inclusion, traffic or payment, and that the preferences do not establish licensing or payment arrangements. Read together with the cost statement, the picture is symmetrical in an uncomfortable way. Restricting grounding has a stated cost of unknown size. Allowing grounding has no guaranteed benefit. Two opposite mistakes follow. One treats the cost sentence as proof of a large loss and decides in a hurry. The other discards it because no number is attached. Both skip the step of holding the statement as a stated cost of unknown size.
The default needs careful wording. Microsoft describes both new preferences as defaulting to allow when no applicable preference is published. That describes the new policy once Microsoft begins honoring preferences, and sites that use the older tags follow a separate transition. It is not a description of how any particular site is treated today, so a team should not report the default as its current setting. What it does mean is that choosing not to decide is itself a position under the new policy, and the record should say so.
A decision under uncertainty is easier to defend, and easier to revisit, when it leaves a record. A short record has six fields, and each has a good and a weak version.
Use. Name one of search discovery, training or grounding. “AI” is too broad to review.
Pages. Describe the pages by type and give examples, such as product explainers or support articles, because the cost may differ by page. Microsoft’s statement does not say which pages bear it.
Choice. State it in the plain terms of the three uses, without configuration detail.
Reason. A strong reason names a business purpose, such as “these explainers exist to answer buyer questions.” A weak reason is “a competitor did it,” which a reviewer cannot test.
Cost named. Use Microsoft’s wording and label it a vendor statement of unspecified size.
Reviewers, review date, and what would change the decision. Name the people, set the date, and write down the triggers.
A reasoned choice can look different at each end, and both can be defensible. A team that allows grounding for its product explainers can say why: those pages exist to answer buyer questions, and Microsoft describes restricting grounding as reducing visibility. A team that restricts grounding for another set of pages can say why too, perhaps because it prefers less reuse of that kind of page and accepts the stated cost. Neither reason has been shown to be correct. Both are easier to defend when the reason is written down and the cost is named.
The review date should follow Microsoft’s calendar. A review set before February 1, 2027 can only revisit the reasoning, because Microsoft describes honoring as starting then. A review set after that date, once the relevant pages have been recrawled, can also draw on observations.
The triggers field is the one most often left blank. Reasonable triggers include Microsoft revising its dates or wording, a change in what the business wants those pages to do, and observations from a baseline recorded before the choice. Writing them down in advance keeps a later reversal from looking like an accident.
The reviewers field matters because a choice about how content may be used can have technical, brand and legal dimensions, and the people who own those should see it before it goes live. This article is not legal advice, and it does not describe how to configure anything. It says only that a decision with a stated cost should not be made by one person on a Friday afternoon.
A stated cost is a reason to decide deliberately. It is not, by itself, a reason to choose one way or the other.
FAQ
What cost does Microsoft state for restricting AI grounding? Microsoft says that restricting grounding reduces brand visibility in AI experiences. The announcement does not give the size of the effect.
Does allowing AI grounding guarantee visibility or traffic? No. Microsoft says neither choice guarantees inclusion, traffic or payment. Allowing grounding is a choice with no guaranteed benefit.
What is the default if a site publishes no preference? Microsoft describes both new preferences as defaulting to allow when no applicable preference is published. That describes the new policy once Microsoft honors preferences, with a separate transition for sites using the older tags, not how any site is treated today.
Does Microsoft’s statement describe ChatGPT, Claude, Gemini, or Perplexity? No. It concerns Microsoft’s products and services. Each of the other engines has its own documentation and controls.
What should a decision record include? The use, the pages, the choice, the reason, the named cost with its source, and the reviewers with a review date and triggers. Labeling the cost as a vendor statement keeps the record accurate.
Is this article reporting a result from Axis Suite? No. It reads a vendor’s published statement and reports no result for any brand.
About Axis Suite Axis Suite is the independent intelligence layer that tracks what AI says about your brand and the observable evidence behind it, retrieval, narrative, recommendation, and corroboration signals, across engines including ChatGPT, Claude, Gemini, and Perplexity. Built by TrendAxis, Axis Suite works from a framework that separates Mentioned, Cited, Recommended, and Chosen. Axis Suite is measurement infrastructure, not a marketing quick fix. Explore the Proof Center.