Semrush is cutting Microsoft Copilot from its Enterprise AIO AI visibility tracking on November 9, 2026, telling customers that changes to Copilot's access requirements block the data it needs. If you've been treating an "AI visibility score" dashboard as the full picture of where your brand gets cited, it just lost a chunk of what it was built on. It won't be the last platform that happens to.
At WebEpex, we already stopped weighting Copilot data in client AI-visibility reports this month, before Semrush made it official. More on why, further down.
What did Semrush actually change?#
Semrush emailed enterprise customers on October 28 saying Copilot tracking inside Enterprise AIO ends November 9, 2026, per PPC Land's reporting. The notice blames "recent changes to Microsoft Copilot's access requirements" without naming one specifically.
The likely culprits: Copilot's August 18 terms requiring a Microsoft account and barring bots and scrapers, plus a sign-in wall Microsoft has been rolling into Edge since October.
Semrush's enterprise plan tracked nine AI surfaces:
- ChatGPT
- Google AI Overview and AI Mode
- Perplexity
- Microsoft Copilot
- Grok
- Gemini
- Claude
- DeepSeek
After November 9, that's eight. Customers get prompt credits back, or they keep the historical Copilot data. Not both. No deadline to choose, either.
Does this affect your AI visibility tracking?#
If you don't pay for an enterprise AI visibility platform, this changes nothing for you today. Read on anyway. You'll likely be buying one of these tools within a year or two.
If you are paying for one, here's the actual problem. The dashboard didn't get worse. It got narrower. And nothing in the product tells you that plainly.
A trend line that drops after November 9 might mean your brand lost visibility. It might just mean Copilot stopped reporting. Those look identical on a chart.
My read: Copilot was never doing the most work for most brands anyway. Similarweb put it at roughly 2% of worldwide generative AI website traffic back in May. But earned-media citations still beat owned content in most AI answers, and 2% of citations isn't nothing. Losing sight of it quietly is worse than losing the traffic itself.
The honest trade-off#
The case for not panicking: Semrush says the other eight platforms are unaffected, and Copilot was the smallest of the nine. If your AI-citation strategy leans on ChatGPT and Gemini, where most volume actually sits, this is a rounding error on your dashboard.
The case for being annoyed: nobody outside these companies can really verify how Semrush, or any AI-visibility vendor, collects this data in the first place. There's no official "did you get cited" API. Every vendor runs some version of send-prompts-and-read-the-answer, and Microsoft just proved that approach is one terms-of-use change away from going dark, for any platform, without warning.
Bing's own first-party tools, like Webmaster Tools AI Performance, only show citations of your own pages. We've already dug into where that leaves you blind. They won't tell you a competitor got picked instead of you. There's no clean replacement for what Copilot tracking did, even a small slice of it.
How we're handling AI visibility reporting at WebEpex#
We stopped building client AEO reporting around any single vendor's visibility score months before this story broke. We treat a tracker's number as one noisy input, never the scoreboard.
We got burned by exactly this pattern once, leaning on one vendor's chart without checking it against a second source, and only caught the gap after a client asked why the line looked worse than reality felt. We don't make that mistake twice.
Every post we ship, on this blog and for clients, carries its own direct answer, FAQs and sourced facts, written so it can be lifted whole by whatever's reading it. We'd rather be the thing that gets cited than the thing measuring whether we got cited.
When a client asks us to prove AI visibility is improving, we don't hand over one chart from one tool. We pull citations across two or three platforms side by side, flag where they disagree, and check Bing's and Google's own first-party citation data against the third-party score. The moment one input vendor changes its terms, the way Microsoft just did to Semrush, a single-source report breaks quietly, and nobody notices until a client asks why the number moved.
We build this kind of AEO and content reporting for clients across the GCC, Europe, USA, Canada and the Indian SMB market, and it's the same posture with every one of them: no dashboard goes to a client without us also knowing what it can't see.
What I'd tell a client asking about this#
If you're on Semrush Enterprise AIO and tracked Copilot, pick a side this week. Bank the returned prompt credits, or keep the historical Copilot data. Semrush isn't defaulting you into either one.
If you don't use an enterprise AI-visibility tool at all, don't rush out to buy one over this news. It's a data-quality story, not a you're-behind story.
Either way: pull your Bing Webmaster Tools citation data once this month and compare it against whatever vendor score you're using. If they tell very different stories, that's the actual finding. Not which number is bigger. We flagged the same kind of blind spot in October's ChatGPT referral data. Trust the raw source data over any one dashboard's read of it.
If you're trying to work out whether your own AI visibility numbers are telling you the truth, send me what you're tracking and I'll tell you straight. Takes two minutes, no pitch attached. cal.com/webepex/growth-review