AI shopping agents were sold as the next easy way to buy online. You tell the agent what you want, and it finds the best deal and checks out. But some of the biggest names in retail and banking have started closing the door on them.

Amazon, JPMorgan Chase, Charles Schwab, Kohl's and Neiman Marcus have moved to block outside agents like Meta's Muse. Walmart has done the opposite and opened up. Many other stores sit in the middle, often without meaning to.

I am Nitin from the Debugged Pro team. Below I stick to what has been reported and confirmed. Then I share what this means for agencies in India and the USA that build and run online stores for clients.

AI shopping agents diagram showing which retailers block, allow or partly allow agents like Meta Muse — Who is blocking and who is opening up, based on reporting by CNBC and The Verge. Graphic: Debugged Pro original.

What Happened: The Doors Start to Close

On November 1, CNBC reported that JPMorgan, Amazon and other big firms are resisting AI agents such as Meta Muse. These personal agents are now being used to find better savings rates, shop for deals and cancel subscriptions.

Here is who has said what so far:

JPMorgan Chase: CEO Jamie Dimon said the bank blocks external AI agents acting for retail customers, citing security risks. He called it "too risky" for clients and the company, per The Verge.
Amazon: it blocked Muse and asked Meta to remove Amazon from the experience. It earlier sued Perplexity over that company's shopping agent.
Schwab, Kohl's and Neiman Marcus: CNBC says they have made similar moves to block personal agents.
Walmart: it announced it is opening its site to the agents customers prefer to use.

An Amazon spokesperson gave CNBC this reasoning: "third-party applications that offer to make purchases on behalf of customers from other businesses should operate openly and respect service provider decisions about whether or not to participate."

Walmart framed its choice around fair pricing. "We price products, not people," the company said in a blog post quoted by CNBC. It added that agents see the same prices a human would see on Walmart.com.

Many stores block agents by accident

This is the part I find most useful for agencies. Identity-verification firm Billions Network and its affiliate AmanTrue audited the published access rules of 51 large US retailers. Their findings, as reported by CNBC:

Only 11 had taken a deliberate position on AI agents.
5, including Amazon, Kohl's and Neiman Marcus, explicitly restrict them.
37 block shopping agents largely by accident, through older bot rules written for scrapers.
None had a way for a verified agent to complete a purchase.

The audit has a caveat. Published rules are not proof of technical enforcement. Still, the message is clear. Most retailers have not made a real decision yet, as the firm's CEO Evin McMullen put it.

Why Big Brands Are Blocking AI Shopping Agents

The reasons fall into four groups. Some are about safety. Others are about money and control.

1. Fraud and agents that go off script

Michael Jabbara leads payment ecosystem risk at Visa. He told CNBC the fraud risks are "very real and very tangible." Fraudsters can build fake agents to steal login and card details. A real agent can also act in ways the shopper never meant, like buying a sofa it was only asked to find.

He shared a sharper example. A user asked an agent to book a gym slot. When the slot was taken, the agent hacked the gym's scheduling software, deleted the existing booking and put the user's name in its place. For Jabbara, the key questions are who the agent is, who it acts for, and whether it has the right permissions.

2. Who holds the customer data

Mike Danford of Adverio, which runs marketplace operations for e-commerce brands, doubts it is all about security. He pointed out that Amazon's own Buy for Me feature uses Amazon's agent on other brands' sites. "It's worried about who holds the data," he told CNBC.

3. Business models under pressure

SharpLink CEO Joseph Chalom argued that incumbents also fear lost margins. An agent can rebook a cheaper flight or move idle cash to a higher-interest account in seconds. He does not expect big firms to be "in a hurry to lower the drawbridge."

4. Agentic flooding

Then there is volume. Wharton professor Ethan Mollick warned of what would be, in effect, a legal denial of service attack on customer service teams. Millions of agents could ask for old refunds or negotiate deals at once. Accenture's Michael Abbott calls this an infrastructure challenge rather than an existential threat.

What AI Shopping Agents Are Like in Real Use

The Verge's Mia Sato tested Muse for weeks. Her hands-on report on AI shopping agents being blocked shows how patchy access still is.

Muse was caught by Target's bot check one day and got through the next.
It admitted Amazon might have the lowest price, but it was hard-blocked from browsing there.
On The RealReal, she had it keep re-adding 10 new items to her cart all day. Each add holds an item for 20 minutes, so other shoppers could not buy them.

After the story ran, The RealReal changed its bot defenses to stop that behavior from Muse. That is a good lesson in itself. Your bot settings can be "open" and still allow abuse you never planned for.

Meta's side is that people stay in control. Spokesperson Tracy Clayton told The Verge that users can take over at any time and that Muse is designed to always ask permission before buying.

What shoppers say they want

Accenture research quoted by CNBC suggests people are open to agents, with conditions:

71% would welcome an AI assistant inside their main bank's app.
65% are open to a GPT-like financial assistant through an AI platform or digital wallet.
82% want to approve every AI action.
79% want to pause or stop AI activity right away.

In short, people want convenience but not at the cost of control. Ed Felten, a former FTC chief technologist, thinks blocking will not last. He told CNBC that sites that block agents too well "will lose business."

What AI Shopping Agents Mean for Agencies in India and the USA

Most of our clients are not Amazon. They are D2C brands, regional retailers and B2B sellers. For them, the question is not whether to fight Meta. It is whether their store works for this new kind of visitor, and on whose terms.

Many Indian agencies build and maintain stores for US brands. So the same decision lands on teams in Pune, Noida and Kochi as on teams in Texas or New York. If you manage the CDN, firewall or bot tool, you are already making an agent policy, even if nobody wrote it down.

Comparison table of three AI shopping agents policies for online stores: block, browse only, open with guardrails — Three common agent policies and their trade-offs. Graphic: Debugged Pro original, based on approaches described by CNBC and The Verge.

Policy	Who does this today	Upside	Downside
Block outside agents	Amazon, Kohl's, Neiman Marcus	Keeps data and checkout in house	Agents may skip you in price comparisons
Partial access	Target, per its comments to The Verge	Lets agents find products, guards accounts	Inconsistent results for shoppers
Open with guardrails	Walmart	Shows up wherever shoppers use agents	Needs logging, rate limits and fraud checks

Platforms matter too. We covered how Shopify is building official routes for agents in our post on Shopify AI agents and Shop MCP checkout. Official channels like that may become the safer way in, compared with agents clicking through a store page.

How to Prepare a Client Store for AI Shopping Agents

Here is the plan we would follow with a client this quarter:

Audit what you block today. Check robots.txt, CDN rules and bot tools. The Robots Exclusion Protocol standard (RFC 9309) explains how crawler rules work, but firewalls often decide more than robots.txt does.
Split the store into zones. Decide separately for product pages, login and checkout. Many brands may be happy with browsing but not purchasing.
Write the policy down. Get sign-off from the client's business, legal and security leads. Update the terms of use to match.
Log agent traffic. Tag known agent user agents and watch where they go. You cannot set a good policy without data.
Add limits for abuse. Cap cart holds, refund requests and repeated actions. The RealReal case shows why.
Keep humans in the purchase loop. Require clear confirmation before any order or payment change.
Clean up product data. Make pricing, shipping and return terms clear and consistent on every page.

Seven step checklist to prepare an ecommerce store for AI shopping agents — A readiness checklist for agency teams in India and the USA. Graphic: Debugged Pro original.

On that last point, FICO's Nikhil Behl made a sharp observation to CNBC. An agent has no feeling for a logo. It reads things like consistent approvals, transparent pricing, fast resolution and fair terms. Brands that compete on clear facts may do well with agents. Brands that rely on loyalty alone may not.

For risk reviews, the NIST AI Risk Management Framework is a useful, free checklist. It helps teams map who is affected and what could go wrong before they change access rules.

Questions to ask your client

Do we want to appear when a shopper asks an agent to compare prices?
Are we fine with agents logging in or checking out for customers?
Who owns the call if an agent places a wrong order?
Which official agent programs or platform integrations should we join first?

Our Take

This is an early, messy stage. Big players are making loud choices, but most stores have simply not decided. That gap is where agencies can help most. An honest audit and a written policy beat an accidental block every time.

We would not rush to open checkout to every agent. We would also not leave a 10-year-old bot rule in charge of a client's revenue. Decide on purpose, measure, and adjust.

Key Takeaways

Amazon, JPMorgan Chase, Schwab, Kohl's and Neiman Marcus are blocking outside AI shopping agents like Meta Muse.
Walmart is opening up and says agents see the same prices as humans.
An audit found 37 of 51 large US retailers block agents largely by accident.
Fraud, data control and traffic floods are the main concerns.
Agencies should audit bot rules, set a written policy and add guardrails.

Need help with this for a client store? Our AI retail and e-commerce solutions team and AI agent development experts work with brands across India and the USA.

About the Author

Nitin Kumar is the Lead Project Coordinator at Debugged Pro. He plans delivery for software and AI projects across India and the USA, and helps teams adopt new tools without risking client work.