Share Check
Three lines, dated, so a reader who stops here still gets the verdict.
- Measure, checked November 17, 2026 on a full read: not stated. The item says share in Wells Fargo's US tracking and nothing more specific. It does not say users, downloads, visits, time spent, or daily or monthly active users.
- Period, checked November 17, 2026: October 2026, in the United States. The item is dated November 16, 2026 and compares October with September on its wider agent figure. No tighter date range is given for the 0.8 percent figure itself.
- Comparison set, checked November 17, 2026: Muse is the only product with a number on this measure. Grok is named as behind Muse, with no figure. ChatGPT, Gemini, Claude, and Meta AI get no share figures in the item at all.
Promise Ledger
- Snippet claim, November 16, 2026: this site's demand-log sweeps at 11:06 and 13:06 CST surfaced a Yahoo Finance / Wells Fargo snippet putting Muse at about 0.8 percent US share, ahead of Grok. Both sweeps left it snippet-only and not banked, on one condition: a full read.
- Full read, this post, November 17, 2026: the item was found and read in full. The 0.8 percent figure is in it, word for word as the snippet carried it. The denominator, the method, and the Grok figure are not in it. Result printed below.
- Still open: what Wells Fargo's tracking counts, and Grok's number on the same measure. Recheck dated November 23, 2026, for a Wells Fargo note or a fuller republication that states either one.
- Launch check, November 17, 2026: Meta opened Muse in Australia on the day this post was checked. The launch gets its own section below, with its launch connectors, its nine named local partners, and its missing AUD prices. It does not change the share verdict. The share figure is US tracking for October 2026, and Australia opened in November, outside the United States.
The item this site actually found
The item is titled "Google and Meta Tell Two Different Stories as Search Rebounds." Yahoo Finance published it on November 16, 2026. The byline is Moz Farooque, and the page states the article first appeared on GuruFocus. It is a short stock-context item, about six paragraphs plus an investors takeaway section, and its main frame is Google search, not Muse. Muse takes one sentence of it, plus two figures in other measures covered below.
That provenance is part of the finding. The snippet travelled as a Wells Fargo tracking result. The full item is a GuruFocus article republished on Yahoo Finance, reporting Wells Fargo figures at second hand. No Wells Fargo note is linked or reproduced, no price target or rating for Meta appears in the item, and the only analyst named anywhere in it is Wells Fargo analyst Ken Gawrelski, named for the sessions figure, not for the share figure.
The Muse sentence, in full, is this: "Muse, Meta's AI agent for tasks such as shopping and booking travel, reached about 0.8% share in Wells Fargo's U.S. tracking, ahead of Grok."
That sentence is the whole share record. Everything after this section is about what the item does not add to it, and about the separate figures it does give.
0.8 percent of what: the item does not say
A share needs a whole. This item never names one for the 0.8 percent figure.
It does not say share of US users. It does not say share of AI agent users, share of downloads, share of visits or web traffic, share of time, or share of tasks. The word tracking is the only method word in the sentence, and the item never defines it: no data provider, no survey, no panel, no app measurement service, no sample size, and no statement of what a tracked unit is.
The sentence before it uses a different, defined measure, and the two are easy to merge by accident. The item reports that AI agents as a group reached 24.4 percent of US users in October 2026. That is a reach figure with a stated whole: US users. Nothing in the item says Muse's 0.8 percent is a slice of that same whole, a slice of the 24.4 percent, or a slice of anything else. A reader who divides one by the other is doing arithmetic the source did not do.
This is the third kind of number problem this site has logged in a week, and it is different from the other two. The Ember fight, settled November 17, was two real columns of one pricing table read as rival prices. The Poppy validator story, November 16, was a real program that turned out to be a command-line tool instead of the web page readers expected. This one is a real figure with its whole missing. The number is not disputed. It is unfinished.
Ahead of Grok, with no Grok number
The comparison in the sentence is verbal only. Ahead of Grok is the complete Grok record in the item: no percentage, no user count, and no statement of how far behind.
No other product gets a share figure either. ChatGPT and Gemini appear in the item, but only in the sessions-per-day comparison below. Claude does not appear in the article body. Meta AI does not appear either, which removes one common confusion at least: the item does not present the 0.8 percent as a slice of Meta AI usage, or Muse as a rename of Meta AI. It describes Muse as Meta's AI agent for shopping and travel booking, and measures it, somehow, on its own.
So the ranking claim cannot be checked from this source. Muse ahead of Grok could describe a gap of a tenth of a point or several points, on a measure nobody reading the item can name. This site is not printing it as a ranking. It is printing it as a claim one republished item made on November 16, 2026, with the evidence it carried, which was none beyond the sentence itself.
The figures in the item that do carry a measure
Three other Muse and agent figures in the item come with enough wording to use, each on its own measure. None of them is the share figure, and none of them fills its gap.
First, reach for the group. AI agents reached 24.4 percent of US users in October 2026, according to the Wells Fargo figures the item reports, up 0.5 percentage points from September, after a gain of 1.25 points the month before. The item's own framing is that growth is slowing. This figure covers AI agents as a category. It is not a Muse figure, and the item does not break it down by product.
Second, frequency for Muse itself. Users averaged 10.8 sessions a day on Muse, roughly two to two and a half sessions more than on ChatGPT or Gemini, a figure the item attributes to Wells Fargo analyst Ken Gawrelski. The item gives no exact ChatGPT or Gemini session numbers, so this post does not print any: subtracting the stated gap from 10.8 would produce figures the source never stated. What the item does support is narrower. On Wells Fargo's count as reported here, Muse users came back more times per day than ChatGPT or Gemini users, in October 2026, by about two to two and a half sessions.
Third, a daily user count outside the US. The item states the app reached about 600,000 daily users outside the United States after launching in Japan and Singapore in late October 2026. No source inside the item is named for this figure beyond the surrounding Wells Fargo context. It is a daily user count, not a share, and it is outside the US, so it cannot be combined with a US tracking percentage whose whole is unstated.
The 10.8 above is sessions per day. It is not 10.8 million users, and it is not the 10 million user figure that has circulated separately in Muse coverage. This item contains no 10 million or 11 million figure at all. Three numbers that look alike, 0.8, 10.8, and 10 million, measure three different things, and only one of them, sessions per day, is defined in the item that carries it.
The Google half of the item
Half the item is about Google, and a reader who only saw the Muse snippet would miss that the share sentence sits inside a stock story. The item reports that growth in Google's traditional search volume picked up by about two percentage points year over year, to 31 percent, while Google continued to lose share, with no share figure given. Wells Fargo credited much of the search rebound to the broad rollout of Google AI Overviews.
For Meta, the item's takeaway is cautious in its own terms. The early numbers suggest Muse is drawing users back often, even as its gains slow. The next test, in the item's framing, is whether Muse keeps users returning as it reaches a wider audience. Completed tasks such as purchases and travel bookings would give a clearer read on commercial potential than return visits alone. That is an investor framing, from a bank's tracking as filtered through a GuruFocus summary. It is not a Meta statement about Muse's users, and Meta has published no user-share figure in any source this site has in its record.
Counts this post keeps separate
Two counts in this site's own record sit near this story and are not evidence for it.
The over 10 million users figure in circulation is a total, not a share, it is not a Meta primary in this record, and it does not appear in the Wells Fargo item. A total cannot supply the missing whole for a percentage. If 0.8 percent were a share of users, the item would still need to say which users, counted where, over what period.
Muse Spark Open, the open model, stood at 14,457 downloads last month, 761 likes, and 10 Spaces in this site's November 17, 2026 sweep reads of its Hugging Face page. That is a download count for a model file. It is not a user count for the Muse app, not a US figure, and not a share of anything. Model downloads and app tracking measure different products in different places.
The September 2026 baseline research this site works from predates any share tracking of this kind and contains no Wells Fargo figure. Nothing in this post rests on it.
Three standing records hold flat beside this check. Muse calling businesses remains paid and US only, on the November 12, 2026 record. Spaces sharing is still not live: Meta described it on November 6 as exploring ways to share. Expedia in Muse remains hotels only, with flights not verified. A share figure with no denominator changes none of those rows.
Australia opened on November 17, after this figure's period
One Muse story did break on November 17, 2026, and a reader who saw it should not have to wonder why it is not the headline here. Meta opened Muse in Australia that day. Superpower Daily reported the launch on November 17. Its reporting rests on Meta's own announcement post, "Muse Is Now Available in Australia: The World's First Personal AI Agent Built for Everyone." This site's news watch filed the launch the same day, on that full read.
At launch, Muse in Australia connects to Gmail, Google Calendar, Instagram, and OpenTable. Nine early local partners are named: Canva, which is also available worldwide, Domain, realestate.com.au, SWEAT, Kogan, MECCA, Officeworks, Ticketek, and ShopBack. Their features roll out in stages after launch. The announcement does not say all nine are live on day one, and this post does not say it either.
Meta published no Australian prices with the launch. Its wording is that Muse is free for most people's needs, with subscriptions for heavier use. No AUD figure is given, and no free tier limit is stated. This post prints neither.
The launch changes none of the figures above, for two dated reasons. The Wells Fargo figure is US tracking, and its period is October 2026. Australia opened on November 17, after that month closed, in a different country. The item's one outside US count, about 600,000 daily users, follows the late October 2026 launches in Japan and Singapore. It predates Australia too.
Australia gets its own full check next on this site: which of the nine partner features are live and which are still staged, and whether Meta publishes AUD prices or free tier limits. Until that check runs, the launch record is the one in this section, dated November 17, 2026, and no partner or price claim beyond it is banked here.
The question for readers
When you see an AI assistant quoted at a small share figure, which measure would actually change what you use: users, time spent, or tasks completed, and why? Answers that name a measure and a decision it would change will shape the November 23 recheck. The Wells Fargo item, read in full, offers frequency, 10.8 sessions a day, as its only defined Muse measure. Frequency is also the measure most readers never see quoted.
Quick answers
What is Muse's US share?
No defined figure exists in the source checked. A Yahoo Finance item dated November 16, 2026, reporting Wells Fargo figures, puts Muse at about 0.8 percent share in Wells Fargo's US tracking. Read in full on November 17, 2026, the item never states what that is a share of.
Is the 0.8 percent a share of US users?
The item does not say so. Its stated US user figure is a different one: AI agents as a group reached 24.4 percent of US users in October 2026. The item does not connect the 0.8 percent to that whole.
What is Grok's share on the same measure?
Not given. The item says Muse is ahead of Grok and prints no Grok figure, so the size of the lead cannot be stated from this source.
Who measured the 0.8 percent, and how?
Wells Fargo, on the item's attribution. The item, a GuruFocus article by Moz Farooque republished on Yahoo Finance, describes no method, data provider, survey, or sample, and names analyst Ken Gawrelski only for the sessions figure.
What Muse figure in the item is defined?
Sessions per day: 10.8 on average for Muse users, roughly two to two and a half more than ChatGPT or Gemini, attributed to Wells Fargo analyst Ken Gawrelski. The item also states about 600,000 daily Muse users outside the US after the late October launches in Japan and Singapore.
Has Meta published a Muse user share?
No Meta primary in this site's record supports any user-share figure. Meta is not quoted on share in the item, and the item gives Meta no share statement of its own.
Frequently Asked Questions
Q: Did the full read debunk the 0.8 percent figure?
A: No. The figure is in the item, in the sentence quoted in this post. What the full read removes is the support around it: no denominator, no method, and no comparison figures. The accurate statement after the full read is that a republished item reported about 0.8 percent in Wells Fargo's US tracking on November 16, 2026, without saying 0.8 percent of what.
Q: Could the 0.8 percent be a share of the 24.4 percent agent reach?
A: The item does not say that, and this post does not assume it. The 24.4 percent is reach among US users for AI agents as a group in October 2026. The 0.8 percent sentence names no whole at all. Joining the two would be this site's arithmetic, not the source's claim.
Q: Why not just call it market share?
A: Because market share names a market, and the item names none. Market share of AI assistant users, of app downloads, and of web visits are three different rankings that can order the same products differently. The item's own wording, share in Wells Fargo's US tracking, is the most specific label the source supports, so this post uses that wording.
Q: Is 10.8 a user count?
A: No. In the item it is sessions per day per user on Muse, attributed to Wells Fargo analyst Ken Gawrelski. The item contains no 10 million or 11 million user figure. Its only user count is about 600,000 daily users outside the US.
Q: Does this change what Muse can do, or what it costs?
A: No connection is stated in the item. It is investor context about usage, not a product announcement. Muse's subscription prices and the Ember developer prices settled in this site's November 17, 2026 post are separate records, and neither rests on a share figure.
Q: When will this site check the share figure again?
A: November 23, 2026, for two things: a Wells Fargo note or fuller republication that states the denominator and method behind the tracking figure, and a Grok figure on the same measure and date. If neither exists by then, this site will say so with that date, and the 0.8 percent will stay labelled as reported, not defined.
Sources
- Yahoo Finance, "Google and Meta Tell Two Different Stories as Search Rebounds," November 16, 2026, by Moz Farooque, first appeared on GuruFocus. Read in full on November 17, 2026. The Muse sentence, quoted in this post, puts Muse at about 0.8 percent share in Wells Fargo's US tracking, ahead of Grok, with no denominator or method stated. Also in the item: AI agents reached 24.4 percent of US users in October 2026, up 0.5 percentage points from September after a 1.25 point gain the month before; Muse users averaged 10.8 sessions a day, roughly two to two and a half more than ChatGPT or Gemini, attributed to Wells Fargo analyst Ken Gawrelski, with no exact ChatGPT or Gemini figures given; about 600,000 daily Muse users outside the US after late October launches in Japan and Singapore; Google traditional search volume growth picked up by about two percentage points year over year to 31 percent, Google continued to lose share with no figure given, and Wells Fargo credited much of the rebound to the broad rollout of AI Overviews. No price target, rating, Grok figure, or 10 million figure appears in the item.
- This site's demand log, November 16, 2026 sweeps at 11:06 and 13:06 CST, and November 17, 2026 sweeps. The November 16 sweeps surfaced the Yahoo Finance / Wells Fargo snippet, about 0.8 percent US share ahead of Grok, as snippet-only and not banked, carried forward on a full read only. The November 17 sweeps through 13:06 CST reported no major new Muse story. The 15:06 CST sweep confirmed the Australia launch below, and the 17:06 and 19:06 CST sweeps report no major Muse story after that launch. The standing records used here hold through the 19:06 sweep: calling paid and US only from November 12, Spaces sharing not live, Expedia in Muse hotels only with flights not verified.
- Superpower Daily, November 17, 2026, published 12:59 PM CST, read in full that day, reporting Meta's about.fb.com post "Muse Is Now Available in Australia: The World's First Personal AI Agent Built for Everyone." Facts used in the Australia section: Muse opened in Australia on November 17, 2026; launch connectors are Gmail, Google Calendar, Instagram, and OpenTable; nine early local partners are named, Canva (also worldwide), Domain, realestate.com.au, SWEAT, Kogan, MECCA, Officeworks, Ticketek, and ShopBack, with partner features rolling out progressively after launch; no AUD pricing is published, and Meta describes Muse as free for most people's needs with subscriptions for heavier use, giving neither prices nor free tier limits. This site's news watch for November 17, 2026 filed the launch at 15:06 CST on that full read.
- This site's Hugging Face reads, November 17, 2026 sweeps: Muse Spark Open at 14,457 downloads last month, 761 likes, and 10 Spaces. Used here only to keep a model download count separate from an app share figure.
- DuckDuckGo checks, November 17, 2026, run for this post. Five queries on the Wells Fargo share snippet, the Yahoo Finance item, and Muse market share. The Yahoo Finance item was found on the second and fifth queries and read in full in the live browser. A direct text fetch of the Yahoo Finance page was blocked by domain policy in this run; the full read rests on the live browser read of the same page, same date.
- Baseline check: this site's September 2026 research file contains no Wells Fargo tracking and no Muse user-share figure. Nothing in this post rests on the baseline file.
Comments