Chase Blocks AI Agents, Bank of America Shifts Risk, Walmart Builds Profiles

Walmart will give every AI agent "a dedicated profile within the customer's account" within a month of 21 October, while JPMorgan Chase refuses to let agents inside at all. Three of America's largest consumer businesses answered the same question in October with three different answers. The question: what to do when a customer sends an AI agent to act on their behalf — buy groceries, book a flight, check a bank balance.

What are the three approaches?

The first is a block. JPMorgan Chase CEO Jamie Dimon said on the Q3 2026 earnings call: "We are not gonna allow just any agent to come, outside agent to come inside the company." The Verge reported that Charles Schwab blocks agents from accounts too, and that Delta prevents them from booking tickets.

The second is a shared login. The agent signs in as the customer, and the customer carries the risk. Bank of America told customers never to share their login with an AI agent that offers to act for them, and warned that if they do, "you're responsible for activity resulting from that access," as TechCrunch reported. The password manager 1Password built a product for this model, 1Password for Agents: an agent asks for a login, the person approves it, and "1Password then fills it directly into the sign-in form instead of sending it to the agent." The release does not say whether the website can tell an agent signed in.

The third is a separate profile. Walmart's Agent Profiles give each agent its own identity inside the customer's account, with approval required for every purchase or "a spending limit they set – for example, $50, $100 or $200." Walmart says agents see the same prices a person sees, and that agents breaking its rules get blocked.

Walmart explained the gap it is closing: "Today, when a customer uses an AI agent, it signs in using a customer's account. This means that all purchases – whether made by the customer or the agent – appear together, without any differentiation."

Why does a separate profile matter?

An AI agent does things the customer would not do, on the customer's behalf. It might browse more pages than the customer would — maybe all of them. Beyond that, the differences could be anything.

That is why a shared login fails as an identity model. Send a dog to shop for you at a local market with a set amount of money; the arrangement only works if the sellers know who sent the dog.

Walmart's spending limit is the set amount of money, and the profile is the sellers knowing who sent it. A block stops the customer from using an agent at all. A shared login hides the agent behind the customer.

How much time do businesses have?

More than most assume. DoorDash CEO Tony Xu said on the Q3 call that outside AI agents have placed "maybe several hundred orders in total," as TechCrunch reported. In the same quarter, DoorDash handled 993 million orders — more than 10 million a day.

The technical rules for how an agent identifies itself to a business remain a draft. The AI customer-service company Sierra and Meta are writing them as the Personal Agent Protocol, with 40-plus companies helping shape it; its website named no business using it with customers as of 1 November. On 30 October, the draft added a recommendation that agents sign their requests with Web Bot Auth, a proposed method for an agent to prove which company runs it. According to the bot-protection company DataDome, as PPC Land reported, Meta's Muse agent has been signing some requests that way since around 12 October, without any announcement.

What should marketers do before agents arrive?

Three practical steps follow from the source:

• Check what the software that keeps bots off your website does to an AI agent today — it might block it, let it through, or be unable to tell it from a person.
• Decide which of the three strategies fits, and write it down, before a platform or a security vendor decides for you.
• Ask whoever runs your website to watch for agents that sign their requests, because those signatures tell you which company operates the agent.

The underlying constraint is simple: if you cannot tell the agent from the customer, you cannot treat it differently. Businesses that wait will get whatever default their bot-protection vendor, their platform, or the arriving agents impose on them.