Your Consumer Just Hired Staff. Now Your Brand Has To Apply.
Amazon blocked Meta's Muse from its store. Shopify shipped Shop Pay into it the next day. They are fighting over a consumer that neither of them owns, and the tables are turning faster than anyone in loyalty expected.
By Jacek Materna, CEO, TrueLoyal | 7 min read | Part 2 of a series
On September 20, Amazon cut off Meta's new Muse agent. Try to shop Amazon through Muse and you hit a wall that reads, in effect, you are not authorized to be here.
On September 21, Shopify put Shop Pay inside Muse.
One day apart. Same product. Two of the largest commerce companies on earth looked at the same thing and made opposite bets. One built a wall. One opened a door.
Here is the bottom line. Both of them are fighting over a consumer that neither of them owns, and the consumer is going to win. Which means every brand now has a question to answer that it has never had to answer before.
Three weeks that collapsed the interface
Meta launched Muse on September 8. In three weeks it passed 3.4 million downloads and took the number one spot on both the App Store and Google Play. Free to start, twenty dollars a month, a hundred for the heavy tier. Meta's stock climbed roughly 30% in a month.
OpenAI launched Dots at DevDay on September 29. Always on, powered by GPT-6 Astra, each one with its own cloud computer and browser, connected to more than four thousand apps.
And Instinct, which is my favorite thing in this whole story, has no app at all. You text it. Fourteen employees, invite only, just raised a billion dollars at a ten billion dollar valuation. A twenty three year old founder and a phone number.
Google shipped Gemini Spark in May. Manus shipped Cue, where every agent gets its own email address, phone number, wallet and computer.
Notice what is not happening here. Nobody is building a better store. Everybody is building a better consumer.
When the agent asks who you are, what does your brand have to show?
TrueLoyal helps consumer brands that sell online and on shelf own the consumer relationship rather than rent it from a platform.
Amazon has 68.6 billion reasons to slam the door
Amazon did $68.6 billion in advertising revenue in 2025, up 22% on the year before. That business depends on a human being landing on an Amazon page and seeing sponsored placements before they choose.
An agent does not see the sponsored placement. It does not scroll the carousel. It reads the brief, finds the item, buys it, and leaves. Every dollar of that ad revenue is leaking out the side of the funnel.
So Amazon has spent a year keeping agents out. It sued Perplexity over Comet. It moved against shopping agents from Google and OpenAI. It won an injunction against Perplexity in March.
Then on August 4 it lost that injunction, and the reason is the most important sentence in this entire story.
The Ninth Circuit held that under federal computer access law, it was Comet's users who accessed Amazon. Not Perplexity.
Read that again. A court looked at an AI agent buying something and decided the agent was the consumer's hands. Not a bot. Not a third party. The person.
Which means every wall built against agents is a wall built against your own customer. Amazon is welcome to make that trade. It has the selection to survive it. Almost nobody else does.
The number that should stop you
Here is the part the download charts hide.
In consumer research around the Muse launch, 8% of Americans said they would trust Meta with their passwords. Google got 30%. Apple 23%. ChatGPT 16%.
So 3.4 million people handed their calendar, their inbox and their shopping to a company that roughly nine in ten of them do not trust. They did it anyway, because the utility was worth more than the discomfort.
That is not loyalty. That is a consumer with no better option, holding their nose.
There is an enormous amount of trust sitting unclaimed in that gap, and platforms are not going to be the ones who claim it.
Part 1: discovery already moved.
Before agents buy anything, something has to name your brand. Why AI search gives a category one winner, and why on-shelf brands are the most exposed.
The tables are turning, and this is the part I find genuinely exciting
For twenty years the arrangement ran one direction. The consumer applied to you. They filled in your form, accepted your terms, handed over their email, and you gave them points. Your program, your rules, your database.
Look at what the consumer is holding now.
An agent with its own email address. Its own phone number. Its own wallet. Its own computer. Instinct has a Trusted Person Network, where your agent talks directly to other people's agents to sort out plans. MCP is quietly becoming the way all of it reaches all of your data.
The consumer is running the program now. They have preferences, standing instructions, a spend ceiling, an approved list and a block list. They have terms.
The question stops being whether a consumer will join your loyalty program. It becomes whether their agent will let you into theirs.
That is a hard flip for a lot of marketing organizations to hear. I think it is the best thing that has happened to this category in a decade.
Which forces a question brands have avoided
Are you loyal to your consumer?
Not the slide. The actual behavior. Do you hold their data carefully. Do you make the return easy when it costs you. Do you price the same whether they are paying attention or not. Do you give them something when there is no purchase attached to it.
Because the second question arrives right behind it, and you do not get to control the answer.
Are they loyal back?
You are about to find out with a precision that did not exist before. An agent will hold the standing instruction. An agent will name the preferred brand or it will not. An agent will accept the substitute or it will refuse. All the ambiguity that let a brand tell itself a comfortable story for twenty years is going to resolve into a yes or a no, at machine speed, thousands of times a day.
The brands that earned it will be in the brief. The brands that bought their way in will find out what they actually bought.
You cannot be in a brief written by a consumer you have never met.
Receipt scanning and rewards connect retail and distributor purchases to known consumers, so brands that sell online and on shelf build first-party data from both.
If you sell on shelf, you are not even in the address book
I said in Part 1 that brands selling through retailers are the most exposed, and it gets worse here rather than better.
An agent works from what it knows about its person. Purchase history. Stated preferences. Past satisfaction. If most of your volume moves through Target and Kroger and a distributor, you do not appear in any of that. You are a product the agent can find, not a brand the consumer named.
Which puts you exactly where you do not want to be when an agent is comparing. In the substitutable pile, being sorted on price.
Knowing who bought, wherever they bought it, stopped being a marketing nicety. It is the entry requirement for the next decade of commerce.
Start here this week
Go text one of these agents and try to reorder your own product. Any of them. Take ten minutes.
Watch three things. Whether it can complete the purchase at all. What it substitutes when it cannot. And whether anything it knows about the person it is working for has your brand's name in it.
That third one is your real score. Everything else is fixable.
Why this is the best moment loyalty has had
For most of my career loyalty was the department that ran the points. Useful, measurable, slightly unfashionable, several rows back from strategy.
Platforms are now spending billions and suing each other over one asset. The consumer relationship. That asset is what loyalty was always, actually about. The whole rest of the business is arriving at the conversation we have been having for twenty years.
The consumer is going to win this fight. They will get the agent, they will get the leverage, and they will start pushing brands to change the terms. The brands that already treated that relationship as something they earned are going to walk into it holding everything they need.
The rest are about to meet a consumer who finally has staff.
I am spending the next run of episodes on this
The Loyalty Recap is my series on LinkedIn. A few minutes an episode, real brands, real numbers, no deck voice. The next stretch stays on this. Personal agents, business agents, who blocks who, and what it does to a brand that never knew its consumer in the first place.
Follow along if this one is keeping you up too.
Follow The Loyalty Recap
My series on LinkedIn. Real brands, real numbers, five minutes or less.
Loyalty for consumer brands that sell online and on shelf.
Schedule a demo and we'll show your Loyalty ROI in 30 minutes or less.
Author
Jacek Materna is the CEO of TrueLoyal and the host of The Loyalty Recap. He writes about loyalty, CPG, and what happens to consumer brands when the shelf stops being a shelf. Connect with him on LinkedIn.
Sources
- Amazon blocking Meta's Muse agent, September 20 to 21, 2026. GeekWire, The Register, Retail Dive, Payments Dive. Amazon cited its Conditions of Use and lack of advance notice.
- Shopify bringing Shop Pay to Muse, September 21, 2026. Forbes, Investopedia, Yahoo Finance. Deutsche Bank described the deal as strategically important.
- Amazon advertising revenue of $68.6B in 2025, up 22% from $56.2B in 2024. Forbes, BigGo Finance, company filings.
- Ninth Circuit ruling, August 4, 2026, holding Amazon unlikely to prove Perplexity's Comet had accessed its computers under federal and California computer-access law, finding that Comet's users accessed Amazon. Reuters, Forbes. Amazon had won a preliminary injunction in March 2026.
- Muse launch September 8, 2026 on iOS, Android, muse.ai and WhatsApp. 3.4M downloads by September 25 per Sensor Tower; Apptopia estimated 4.3M; Appfigures estimated 2.3M. Number one on US App Store September 18 and Google Play September 19. TechCrunch, CNBC, The AI Insider.
- Consumer trust research around the Muse launch. 8% of Americans said they would trust Meta with their passwords, against 30% Google, 23% Apple, 16% ChatGPT. Reported by Yahoo Tech and others.
- OpenAI Dots launched at DevDay, September 29, 2026. Powered by GPT-6 Astra, each agent with its own cloud computer and browser, 4,000+ app connections, rolling out to Pro, Business Premium and Enterprise. OpenAI announcement, TechCrunch, Forbes, Quartz.
- Instinct, built by Spear Street Technology, founder Noah Shinn. Invite-only since August 2026, SMS, iMessage, WhatsApp and voice, no mobile app, 14 employees, 100,000+ users. Raising approximately $1B at a $10B valuation. Trusted Person Network allows agent-to-agent coordination. Business Insider, The Tech Portal.
- Gemini Spark, Google, May 2026. Manus Cue, providing each agent a dedicated email, phone number, wallet and cloud computer. Forbes, industry reporting.
