Home Depot will let AI agents buy for customers, but only if the agents prove they are not bots and prove whom they represent.
That is the position of Angie Brown, Home Depot’s executive vice president and chief information officer, in a Business Insider interview published Oct. 28 and republished by AOL. I have tracked this fight at Ecommerce Paradise since Meta launched its Muse agent, and Home Depot has now said out loud what it wants from software that shops on a customer’s behalf. If you sell furniture, outdoor gear, appliances, or any other big-ticket item, the identity standard that retailers settle on is the one your store will be measured against.
Here is what Brown said, how Amazon and Walmart got to their own rules, what it does to a store with $1,500 to $5,000 orders, and five things to do before the holiday rush.
When software starts placing your big orders, a bad one becomes a chargeback, not a return. ClearSale sells fraud and chargeback protection for online stores, which is the layer that matters most once agents are buying. Add fraud and chargeback protection with ClearSale →
Home Depot Will Let AI Agents Buy, But Only With ID
Brown told Business Insider that Home Depot wants to support agents that search for products, compare them, and buy on a customer’s behalf. The condition is verification. “We want to know you’re not a bot, and we actually want to know if it’s you,” she said, per the AOL republication of the interview.
Home Depot is in active talks with Meta about how Muse, Meta’s personal AI agent, should identify itself to retail websites. “There’s openness on both sides, because they want it to grow,” Brown said.
She also said Home Depot does not want to lose its direct relationship with loyal customers, such as the contractors enrolled in its Pro Rewards program, according to the same interview.
Per PYMNTS, which paired the Home Depot comments with Walmart’s agent rules in a piece published Oct. 29, Home Depot has not built a gateway for agents the way Walmart has. It asks only that agents identify themselves and confirm whom they represent.
What the coverage I found does not include: a launch date, a technical spec, or a list of approved agents. Home Depot is describing a direction. Nothing is live.
Amazon Blocked Muse and Walmart Set Caps: The Agent Fight So Far
Meta released Muse on Sept. 8. Twelve days later, on Sept. 20, Amazon said Muse could no longer buy on Amazon.com, per GeekWire reporting summarized by PPC Land. Amazon listed four objections: it never agreed to Muse operating on its store, Meta did not tell Amazon that Muse would reach it, Muse does not identify itself when browsing, and Muse appears to capture and store customer credentials.
An Amazon spokesperson said third-party shopping agents “should operate openly and respect service provider decisions about whether or not to participate.” Meta has said Muse “has no visibility into people’s passwords or payment methods.”
Walmart went the other way on Oct. 21. Per TechCrunch, Walmart introduced an Agent Gateway, a Model Context Protocol server that lets approved agents search its inventory and buy, plus Agent Profiles, which give each agent its own identity inside a shopper’s account. Walmart compared the setup to adding an authorized user on a credit card.
Digital Commerce 360 reported that shoppers can approve every purchase or set a spending cap such as $50, $100 or $200. Agents cannot receive payment-card details, and agents that break the terms get blocked. Unintended agent purchases fall under Walmart’s existing returns process. Profiles launch within about a month, while the gateway starts in limited beta with no announced timing.
Daniel Danker, Walmart’s executive vice president of AI acceleration, product and design, said agents can “shop using dedicated profiles linked to a customer’s account.” Manish Joneja, senior vice president of Walmart Global Marketplace, said agentic commerce “could create more opportunities for sellers to put the right products in front of the right customers.”
Meta and Sierra announced a Personal Agent Protocol on Oct. 6 to standardize how agents introduce themselves, and a draft published Oct. 9 does not yet cover payments, per PYMNTS. A merchant survey cited in the same PYMNTS piece found that 80% of 60 U.S. merchants with $50 million to $1 billion in revenue expect to pay a fee or commission on agent sales. Fewer than 3 in 10 would offer their full catalog to agents on the same terms as their own site. A separate PYMNTS Intelligence report said only 15% of merchants have product data structured well enough for agents to read.
The counterpoint is that none of this is settled. Home Depot is talking, not shipping. The protocol draft skips payments. And the three retailers in this story are doing three different things: Amazon blocks, Walmart builds a gateway, Home Depot asks for ID first. PYMNTS CEO Karen Webster wrote in an Oct. 28 column that “Walmart decides who gets in, what they see and what they spend,” which is a fair description of who holds the power in each model.
I covered the opening moves in Walmart Opens to AI Shopping Agents as Amazon Blocks Them. On the Meta side, Muse passing 10 million users with shopping as its top use is why retailers are negotiating now.
What Home Depot’s ID Rule Means for High-Ticket Store Owners
My read: Home Depot’s approach is the one that translates to independent stores. Walmart can build a gateway and Amazon can build a wall. You cannot do either. What you can do is decide who gets in, based on whether they identify themselves, and you can do it with the tools you already run.
Here is why this lands harder on high-ticket than on low-ticket. A $40 phone case bought by the wrong agent is a shrug. A $2,400 patio set bought by an agent that misread a request is a dispute, and a dispute is a problem for your processor’s numbers whether or not the customer was happy to keep the product. Walmart’s answer, per Digital Commerce 360, is to route unintended agent purchases through its normal returns process. Most independent stores have not written that rule down anywhere. If a buyer’s card issuer hears the story first, you are in the chargeback fight and not the return flow.
Rough math, and this is hypothetical: say your store does 300 orders a month at a $2,000 average order, which is $600,000 in revenue. If agents place 2% of those orders, that is six orders and $12,000. If one in ten of those agent orders ends in a dispute, you are looking at a dispute every month or two, each one around $2,000 plus fees plus the margin on a product you may not get back. Small. But you have no data on your own agent rate yet, and the first number you need is that one.
Now the scenarios, with thresholds that are mine and not an industry standard. If agents are under 1% of your orders by the end of Q1, do the cheap prep below and move on. Between 1% and 5%, add order verification and a written agent policy. Above 5%, treat agents as a sales channel with its own rules, its own reporting, and possibly its own fee structure.
The Pro Rewards comment deserves more attention than it got. Home Depot’s worry is that an agent sits between the store and the contractor who buys every week. For you the same risk is sharper, because you do not have a loyalty program or a brand people type in from memory. The customer relationship is your phone number, your email list, and your post-purchase experience. I wrote about turning leads into clients in my high-ticket sales follow-up guide, and the logic applies double when the first touch might be software.
Then there is product data. PYMNTS Intelligence puts agent-ready product data at 15% of merchants. High-ticket stores that already run Google Shopping tend to be ahead, because you cannot rank without clean titles, GTINs, and specs. If you have been sloppy there, agents will punish you faster than Google does. Google’s Merchant API added a warranty field this week, which is exactly the kind of detail a buyer’s agent will want to compare. My guide to turning Google Shopping clicks into sales covers the feed side.
One more point of honesty. Muse is one consumer agent. A study I covered found AI shopping picks big retailers nine times in ten, so the near-term risk for a small store is not that agents flood you with orders. It is that they send nearly nothing your way unless your data, reviews, and policies make you the safe pick. Both problems have the same fix.
If you would rather not wire the feed, the phones, the checkout rules, and the fraud screening yourself, that is the kind of setup my team handles in the turnkey done-for-you store build. It is a lot to carry alone while also running ads.
Want to work through what AI shopping agents mean for your store with other store owners and me inside the community? Start the 7-day free trial. Join the Skool community →
Five Steps to Get Your Store Ready for AI Shopping Agents
None of this needs a developer. Do these in order over the next two weeks:
- Find out who is already visiting. Check your bot and firewall settings on your host or CDN, and your robots.txt file, and write down which automated visitors you allow. I would start by logging agent traffic, not blocking it. Amazon can afford a wall and you cannot. Shopify’s fall edition fraud rules and AI shopping controls are worth a read before you touch anything. Your Shopify admin is the first place to look for what you can already switch on.
- Clean your product data. Fill in GTIN or MPN, brand, dimensions, shipping weight, and warranty terms for every product, in Shopify and in Merchant Center. If you sell the same item through several channels, Shopify’s channel mapping lets each channel carry its own product titles, which helps when an agent reads your listing differently than Google does.
- Add a human check on large orders. Pick a dollar line, say $1,000, and call or text the buyer before shipping anything above it. A business phone line from Quo handles calls and texts in one place, and ClearSale can screen the order first. A two-minute call kills most “I never asked for that” disputes before they start.
- Write an agent clause into your terms and returns page. Say what happens when an order is placed through a third-party AI assistant, who approves it, and how returns work, using Walmart’s wording as a model. Have a lawyer review it, because I am not a legal advisor and this is a reported-news summary. Your backup payment plan belongs in the same review.
- Own the customer after the sale. Set up a post-purchase flow in Klaviyo that asks for the buyer’s own email and phone. Then put a help widget from Tidio where agent-related questions will land. This is the Pro Rewards lesson: the store that keeps the relationship keeps the repeat order.
Want a second set of eyes on your setup? Grab a spot through Turnkey service details.
Frequently Asked Questions
Has Home Depot launched AI agent shopping?
No. Angie Brown described a policy direction and talks with Meta, and the coverage gives no launch date. Treat it as a signal and not a switch.
What is Meta Muse and why does it matter to my store?
Muse is Meta’s personal AI agent, released Sept. 8, that can shop, fill forms and buy. My post on Muse checking out with Shop Pay shows how close it already is to real Shopify checkouts.
Why did Amazon block Muse?
Per PPC Land’s summary of GeekWire, Amazon said it never agreed to Muse operating on its store, Meta did not warn it, Muse does not identify itself, and Muse appears to store customer credentials.
Will agent orders raise my chargebacks?
Nobody has published data yet, so I cannot say. Walmart sends mistaken agent buys through returns, and I would copy that approach. For context, service disputes made up 24% of contested chargebacks in one study I covered.
Should I block AI agents from my store?
Not blindly. Log first, then decide. A blanket block costs you nothing today and may cost you the channel later.
Do banks treat agents differently too?
Yes. Citi and BofA restricted AI agents and Canadian banks followed, so expect payment-side friction as well.
Which high-ticket niches are most exposed?
Any niche with spec-heavy products that buyers compare online. My free high-ticket niches list is a good place to see which categories fit that description. The high-ticket dropshipping guide explains the model.
Want my team to build and run your high-ticket store, with phone sales, clean product data and fraud screening in place? See the turnkey done-for-you service →
I will keep watching how Home Depot, Walmart, Amazon and Meta settle the ID question, because it decides who gets to sell to the next wave of buyers. Subscribe to the YouTube channel for daily breakdowns. More breaking news coming through the day.
Related Articles
If this was useful, these go deeper:
- Meta’s Muse AI Can Now Check Out With Shop Pay
- BofA Tells Customers Not to Give AI Agents Passwords
- Newegg Turns On Google Checkout in Gemini and AI Mode
- Claude Duet Can Buy Supplies via Link and Shop Pay
- High-Ticket Niches List: 150+ Best Dropshipping Niches for 2026
Trevor Fenner is an ecommerce entrepreneur and world traveler from Seattle, Washington, and the founder of Ecommerce Paradise, a platform focused on helping entrepreneurs build and scale profitable high-ticket ecommerce and dropshipping businesses. With over a decade of hands-on experience, Trevor specializes in high-ticket dropshipping strategy, niche and product selection, supplier recruiting and onboarding, Google & Bing Shopping ads, ecommerce SEO, and systems-driven automation and scaling. Through Ecommerce Paradise, he provides free education via in-depth guides like How to Start High-Ticket Dropshipping, advanced training through the High-Ticket Dropshipping Masterclass, and fully done-for-you turnkey ecommerce services for entrepreneurs who want a faster, more hands-off path to growth. Trevor is known for emphasizing sustainable, real-world ecommerce models over hype-driven tactics, helping store owners build scalable, sellable, and location-independent brands.
He maintains homes in Seattle and Bali, spends most of the year overseas, and usually returns to Seattle for about a month each summer.
