Only 6% are very comfortable letting AI buy at $500. Meta says Muse shoppers already average $200 and up.
A survey of feelings and a company’s own usage claim point in different directions. Both tell a brand the same thing: agents are already buying, and the buyer is checking what they read about you.
Two numbers came out this week. One says shoppers do not want AI spending real money. The other says some of them already are.
The survey: comfort falls as the price rises
On Oct 20, Wildfire Systems published a survey of 1,007 US consumers. Wildfire describes itself as a loyalty and consumer insights platform. It runs cashback and rewards programs, so it has a stake in how shoppers feel about paying through AI. The release is titled “Agentic Commerce’s Wallet Problem”.
Wildfire says 29% of respondents are “very comfortable” letting an agent make a purchase under $20. For purchases of $500 or more, that falls to 6%. By category, 23% are very comfortable delegating toiletries or household basics, compared with 17% for airfare or hotels.
Who should handle the payment
The same survey asked who people would trust to process an agent’s payment. Only 5% would trust an AI platform itself, and 41% would trust none of the options tested. Banks were the most trusted named option at 17%, then digital wallets at 14%, and credit card networks and major retailers at 12% each.
Wildfire also cites Mastercard’s Signals research: 85% of consumers are open to working with an AI agent to find the best option, while only one in 10 would let an agent complete a purchase on its own. That figure is Mastercard’s, reported inside Wildfire’s release.
The other number: Meta’s Muse shoppers
On Meta’s third-quarter call on Wednesday, Oct 21, Mark Zuckerberg said Muse has been adopted by more than 10 million users, according to TechCrunch. He also said that over a quarter of Muse’s engaged users who were shopping through the AI agent had an average purchase size of over $200.
TechCrunch also reports that about 25% of people who set up Muse now use it daily, and that Muse uses Plaid’s connector to enable agentic shopping. Zuckerberg said Plaid had not seen activity like it on any other service before.
Why they can both be true
One is a survey of how people say they feel. The other is a claim from a company about what its most active users did. They are not the same people, and the second has no published breakdown. Meta sells Muse, so the number is also a pitch. We have only what was said on the call, as TechCrunch reported it.
Read side by side, they still suggest something useful. Shoppers who hand buying to an agent are not only buying paper towels. And Wildfire’s own conclusion is that the near-term opportunity builds from lower-cost, routine purchases, while the bigger fight is over who earns trust when an agent is ready to pay.
What it means for your product pages
The higher the price, the more the shopper will check before saying yes. The agent reads your price, your stock, your shipping promise and your return policy, and the shopper sees them on the approval screen. A wrong or stale detail on a $200 order is a reason to stop.
This is our reading, not a finding from either source. But it points the same way as the data: the more money is at stake, the more your facts need to be right and easy to read.
What to do this week
- Ask ChatGPT, Gemini and Google AI for your best-selling product over $100. Write down the price, shipping time and return window each one states.
- Compare those to your live page. Fix any that differ.
- Put price, stock and return policy in plain text on the page, not only in images or pop-ups.
- Check that your checkout shows the same price the page shows.
How Shopthru.ai helps
Shopthru.ai is an AI agents team for consumer brands. It asks ChatGPT, Gemini, Google AI, Perplexity and Copilot what your shoppers ask, finds where AI gets your products, prices or policies wrong, and drafts each fix into an approval inbox. It ships only on your yes, then checks again. Instagram and Amazon are next.