Log entry Nº 716 · the stack watch
Bearing S 180° · November 13, 2026 · 5 min read
Commerce TechS 180°Watch log
United Airlines blocks unauthorized customer AI agents
United Airlines, the $59 billion carrier, moved Wednesday to block unauthorized customer AI agents across booking and ticketing channels, while Walmart and Target used different access models.
By Amara Osei5 min read910 words
$59 billion United Airlines, the world’s second-largest airline, moved on Wednesday to restrict unauthorized customer AI agents, extending a corporate blocking wave that already includes Walmart, Target and Amazon.
The policy does not shut out every bot. United’s wording leaves room for authorized agents, and the airline says it will keep evaluating how personal AI assistants interact with its services as they evolve. United signaled that it could reconsider the ban as controls change.
United pointed to similar actions by banking and retail leaders and urged others to protect customer data, security and safety.
What exactly did United change?
United said: “United is taking steps to restrict access – directly within our digital platforms and indirectly via travel agencies and aggregators – to prohibit use by unauthorized, third-party personal AI agents or assistants. This work includes updates to our booking, ticketing and distribution policies.”
United cited data leakage and credential sharing, warning that personal assistants can expose sensitive information. Agent actions can trigger unintended purchases, account changes or use of MileagePlus miles. Automated actions may also breach U.S. Federal Aviation Administration and Department of Transportation rules, including flight check-in requirements for human acknowledgement that bags contain no banned substances such as explosives.
Beyond compliance, industry observers saw a larger commercial problem: AI lacks industry-wide standardization even as major vendors and startups add bots. Supporting each bot could create an almost infinite drain on IT resources. They said strictly enforced standards could persuade some enterprises to reconsider their resistance.
Dickson Research principal analyst Frank Dickson said: “Blocking does not destroy the value of agents. It moves it. Every wall eventually gets a gate. The gate charges admission.”
Dickson expects authorized access to return through enterprise-controlled channels. Whoever owns the agent owns the customer relationship, he said, and airlines spent years pulling customers back from third-party travel sites rather than surrendering them to another company’s chatbot.
Takeaway: United has drawn a line around unauthorized agents while keeping the door open for partners it can control.
How are retailers separating bots from people?\n
Because agents usually sign in with a customer’s credentials, bot activity can look identical to human activity. Analysts called separate bot credentials the cleanest answer.
Walmart moved close to that goal in an announcement last month, saying a program scheduled to start this month would distinguish bot and human actions. Agent Profiles would launch in the next month: each customer could authorize an agent with a dedicated profile and see agent purchases separately. Walmart said today’s agent and customer purchases instead appear together without differentiation.
Target took a different route in recent terms-and-conditions changes. It permits only its AI Shopping Assistant, OAI, Google and Microsoft—a four-service allowlist that Target can add to, remove from or change at any time in its sole discretion.
Target says every other agentic-commerce tool is unauthorized and expressly prohibited, whether it accesses the site directly or through APIs, logs in, scrapes, queries, uses automated customer credentials or uses another method. Approved agents cannot operate alongside unauthorized ones unless Target grants explicit written approval.
Takeaway: Walmart separates authorized bots with profiles, while Target controls access through a narrow written allowlist.
Does blocking weaken agent commerce?
Consultants remain divided on whether broad blocks will dilute autonomous agents that make purchases, order plane tickets or deliver concierge functions, even as analysts expect more enterprises to close external bot access.
Dickson said enterprises fear unassigned liability rather than bots themselves. Rules that specify which steps humans or agents may perform, and how each action gets recorded, let general counsel approve deployment; ambiguity keeps bots stuck in pilot programs.
He also warned that a sea of agents creates a sea of requests that somebody must pay for. Each brings its own login method, reading of customer approval and record of what it did, if it keeps one at all.
An airline cannot build a separate door for every agent, he said. Bots that check prices repeatedly do not buy more seats, making a wait for the market to settle reasonable.
Digital 520 principal consultant Noah Kenney saw no downside to United’s ban because the carrier faces substantial risk, uncertain commercial upside and almost no competitive pressure to accept either. He expects United to wait until the legal consequences of bot behavior become more predictable.
Takeaway: Near-term restrictions may slow autonomous purchasing, but they also push enterprises toward accountable, controlled access.
What controls will unlock wider access?
Kenney said the biggest mistake in enterprise AI deployments gives agents access to more information and systems than their responsibilities require. He draws a fundamental distinction between an agent instructed to use one system and one technically unable to reach anything else. Broad capability plus hoped-for restraint remains unsafe.
Acceligence CEO Justin Greis expects more companies to restrict bots while they establish controls. He warned that widespread blocking would reduce agents’ value by preventing them from completing tasks across services.
Greis also wants customer training around authorized actions. People can create obligations through an agent even when they never personally click buy, send or confirm, he said. He expects a cultural shift comparable to how people say things online that they would not say face to face, creating distance between a choice and personal responsibility.
Takeaway: Practitioners will win broader access when permissions are narrow, attributable and technically enforced, not when agents merely promise restraint.
Plot points
- $59 billion United Airlines, the world’s second-largest airline, moved Wednesday to restrict unauthorized customer AI agents.
- United’s policy work covers three named areas—booking, ticketing and distribution—and extends indirectly through travel agencies and aggregators.
- Walmart said last month that Agent Profiles would launch in the next month, giving each authorized agent a dedicated account profile.
- Target allows four named services—its AI Shopping Assistant, OAI, Google and Microsoft—and can revise the list at any time.
- United cited FAA and DOT flight check-in rules requiring human acknowledgement about banned substances in bags.
via cio.com (Source)
More from Amara Osei
Show full bio
Market editor covering technology and product news in marketing technology at Martech Beacon.
47 articles
Neighbouring fixes
- AI Assistants' Rogue Acts Are Chilling Consumer Adoption
- Google Adds New Universal Commerce Protocol Features, AI Retailer Tools
- Amazon Ads unBoxed 2026 Signals Shift to Outcome-Based Buying
- Spyware in 120,000 Sites: Why Apple's ID Block Isn't the Villain Story
- Databricks' agentic CDP integrates IAS to clean first-party data