NewsFor: Customer support · IT and service desk

United Airlines blocks customers' AI agents; Target and Walmart set limits

United, Target and Walmart each set a different rule for AI assistants that log in as a customer. Support and IT can block, approve a list or separate logins.

United Airlines said on 11 November 2026 that it is acting "to prohibit use by unauthorized, third-party personal AI agents or assistants" for booking and managing flights, "directly within our digital platforms and indirectly via travel agencies and aggregators", according to its statement as quoted by [CIO](https://www.cio.com/article/4248161/united-airlines-becomes-the-latest-enterprise-to-shut-its-door-to-customer-bots-2.html) and [TechCrunch](https://techcrunch.com/2026/11/12/united-airlines-blocks-ai-agents-from-booking-flights-citing-cybersecurity-risks/). A personal AI agent here is an assistant a customer lets into their account, usually by giving it their password, so that it can act for them. CIO reports the move follows Walmart, Target and Amazon; it gives detail on the first two, and the three rules differ.

Three companies, three rules

  • United: block. Its statement says sharing credentials with such assistants "can expose sensitive information", that their actions "may result in unintended purchases, account changes, or use of MileagePlus miles", and that automated actions "have the potential to violate FAA and DOT regulations", such as the rule that a person confirms at check-in that bags hold nothing banned. United says it is updating its "booking, ticketing and distribution policies" and "will continue to evaluate" as assistants change.
  • Target: an approved list. CIO reports that Target's terms now allow only its own AI Shopping Assistant and agents from three companies it lists as "OAI, Google and Microsoft". All others "are unauthorized, expressly prohibited", including through "automated use of user credentials", and Target may change the list "at any time and in its sole discretion".
  • Walmart: a separate identity. Walmart said last month that an agent today "signs in using a customer's account", so its purchases and the customer's "appear together, without any differentiation". Its Agent Profiles, due this month, let a customer authorise each agent with "a dedicated profile" and see that agent's purchases separately, CIO reports.

Why this lands on support and the service desk

An agent signed in with a customer's password is indistinguishable from the customer, CIO notes. Wren's read: when it buys the wrong item or changes a booking, the dispute arrives in the support queue with nothing to show who acted. Frank Dickson of Dickson Research told CIO: "Enterprises do not fear bots. They fear liability that nobody has assigned." Each agent, he said, arrives "with its own way of logging in, its own idea of what the customer approved, and its own record of what it did, if it keeps one at all".

The same risk runs the other way, through your own staff. Bank of America last week urged customers not to share passwords with AI assistants, TechCrunch reports. An employee who gives an assistant the login to a travel, supplier or bank portal is doing what these companies now prohibit, and United's block reaches travel agencies and aggregators too.

What the three companies did not say

Neither report carries a number: the share of traffic that comes from agents, how many unintended purchases or disputes prompted the rules, or what any of this costs to enforce. United does not say when the block starts, how it will tell an assistant from a person, what happens to a customer whose assistant is caught, or what would make an agent "authorized". Nothing is said about support staffing or training.

What to do about customers' AI agents this month

  • If you run support: ask whoever owns your terms of use whether they cover a customer's AI assistant using the account, and start tagging tickets where a customer says an assistant made the purchase or change. You need that count before choosing between a block, an approved list and separate agent access.
  • Write down who pays when a customer's assistant makes a purchase the customer did not intend, and give the rule to your agents before the first dispute.
  • If you run IT: add one line to the acceptable-use policy this week, that no company, supplier-portal or bank password is given to a personal AI assistant, and ask your travel agency or booking tool whether United's policy change affects how it books.
Was this useful for your work?

Take this to work

Forward this

United Airlines said on 11 November 2026 that it will block unauthorised third-party AI assistants from booking and managing flights, after Target limited its site to an approved list of agents and Walmart announced separate agent profiles, CIO and TechCrunch report. We should pick a rule for our own customer accounts and tell staff not to give AI assistants passwords to travel, supplier or bank portals.

Questions to ask

  1. How many logins, orders or tickets a month already come from a customer's AI assistant, and can our systems tell them from the customer?
  2. Do our terms of use say whether a customer may let an AI assistant use their account, and who pays when it makes a purchase they did not intend?
  3. Can our support platform or identity provider give an agent its own access, separate from the customer's password, and what does that cost?
  4. Does our travel agency or booking tool use automation that United's policy change would catch?
  5. Which staff already let an AI assistant sign in to supplier, bank or travel portals with company credentials?

Questions this article answers

Does United Airlines allow AI agents to book flights?

United said on 11 November 2026 that it is restricting access to prohibit unauthorised third-party personal AI agents or assistants, on its own platforms and through travel agencies and aggregators, CIO and TechCrunch report. It has not said when the block starts or what an authorised agent would be.

Which AI agents does Target allow on its site?

CIO reports that Target's terms allow its own AI Shopping Assistant and agents from three companies listed as OAI, Google and Microsoft, and prohibit all others. Target says it may change the list at any time.

What are Walmart Agent Profiles?

Walmart says Agent Profiles let a customer authorise each AI agent with a dedicated profile inside their account, so the agent's purchases show separately from the customer's own. CIO reports the launch is scheduled for November 2026.

Why are companies blocking customers' AI assistants?

United cites exposed credentials, unintended purchases and account changes, and federal rules that require a person to complete check-in. Analysts quoted by CIO add that nobody has assigned liability for what an agent does and that supporting many different agents costs IT time.

Send this to someone who runs a teamLinkedInXBlueskyHNEmail