Field report W-85 / M5 · sheet /strategy
Corporate StrategyFiled from the territory
AI Agents Are Eroding the Moats Built on Consumer Laziness
Muse sent Schwab down 6% and Planet Fitness down 11%. The selloff marks a repricing of businesses whose moats rest on customer inertia rather than real advantage.
- Grid
- W-85 / M5
- Surveyed
- November 17, 2026
- Reading
- 5 min read
- Correspondent
- Olivia Hart

Field notes
- Meta's Muse launch in September drove Charles Schwab down over 6%, Allstate down 5.5% and Planet Fitness down as much as 11%.
- A 2025 American Economic Review study estimated cancellation friction roughly doubled subscription seller revenues on average.
- Walmart opened an Agent Gateway to outside AI agents; Amazon blocked Muse and is building its own tools.
- Schwab CEO Rick Wurster said the bank is blocking unauthorized AI agents attempting to move customer money.
Charles Schwab fell more than 6% in a single September session after Meta launched its Muse AI agent — and investors began repricing a question that has quietly underpinned whole industries: what happens to businesses whose moat is human inertia?
Allstate dropped 5.5% in the same selloff. Planet Fitness fell as much as 11%, and travel names including Expedia and Booking Holdings came under pressure, Bloomberg reported. The trigger was not earnings or guidance. It was the prospect of consumers armed with tireless personal AI assistants that compare, switch and cancel on their behalf.
The reaction may prove overdone. Agentic AI remains young, and major questions persist around security, trust and how much authority consumers will actually hand to software. But the underlying test is worth running now, as Fortune commentary published November 12 argues: would a company's advantage still exist if every customer had a tireless assistant constantly hunting for a better deal?
What are 'behavioral moats' — and why are they at risk?
The strongest moats come from something genuinely hard to replicate: unique products, valuable intellectual property, scarce assets or scale. But some businesses owe part of their protection to the fact that their customers are human — busy, forgetful, and unwilling to read the fine print.
We postpone switching banks. We keep paying for subscriptions we barely use. When shopping becomes complicated, we settle for good enough. Call these behavioral moats. Personal AI agents threaten each of them, because software does not get frustrated after the sixth page of fine print and does not forget to cancel.
When switching is the moat
Financial services show the pattern most clearly. Consumers routinely leave money in accounts paying less interest than alternatives offer, because moving means comparing banks, filling forms, changing direct deposits and reconnecting automatic payments. An AI agent faces none of that friction; in theory, it could compare savings rates or brokerage products continuously and eventually handle the switch itself.
Banks have noticed. Schwab CEO Rick Wurster recently said the company is blocking unauthorized outside AI agents attempting to access customer accounts or move money. Schwab is instead building controlled connections for outside AI systems while limiting what they can do, and other major banks are taking similar precautions, according to Bloomberg.
There are legitimate security reasons here. Giving software free access to your life savings is very different from asking it to order paper towels. But friction protects a business whether it exists for good reasons or not — and as agents get better at tedious financial tasks safely, some of that protection weakens.
When complexity is the moat
Sometimes customers stay not because switching is hard, but because comparing is. Insurance policies differ in deductibles, coverage limits and exclusions beyond the headline premium; telecom plans mix data caps, bundles, promotional prices and fees. An agent absorbs all of it without fatigue.
Some insurers are inviting the change rather than fighting it. Kin Insurance said it would welcome AI shopping agents rather than block them, and has built tools that make it easier for agents to obtain home insurance quotes. Marketplace Insurify launched an AI Agent Gateway that lets agents gather and compare real-time quotes while preserving deductibles, limits and disclosures. Being the company an agent can shop most easily could itself become an advantage.
When forgetting to cancel is the moat
Subscriptions are the clearest case. A 2025 study in the American Economic Review found cancellations rose sharply when replacement credit cards forced subscribers to make an active renewal decision. The researchers estimated that cancellation friction roughly doubled seller revenues on average, holding the starting subscriber base constant.
An agent that scans recurring charges, compares them with usage and asks whether you still want to pay could make that inattention rare. That concern helped send Planet Fitness sharply lower after Muse launched, as investors weighed what easier cancellation means for membership businesses.
Netflix is the more interesting case. Some subscribers may stay only because canceling is one more undone chore. But Netflix also owns content people actually want to watch — and no agent can shop around and recreate an exclusive show elsewhere. That is the difference between a behavioral moat and a real one.
When settling for good enough is the moat
Even where prices are clear and switching is easy, comparison takes time. Nobody checks every travel site each morning. Priceline's Penny agent can now compare destinations, prices and availability, and move from a broad request to an actual booking within one conversation.
The sharpest contrast sits in retail. Walmart introduced an Agent Gateway that lets outside AI agents connect directly to its systems, search products and make purchases, with shopper-set limits on what agents can buy. Amazon, by contrast, has blocked Muse from shopping on its platform, citing security and unauthorized-access concerns, and is building its own AI shopping tools instead, Axios reported.
The stakes are the customer relationship itself. Today a shopper goes directly to Amazon or Walmart for a television. Tomorrow that shopper may tell an agent, "Find me the best 55-inch television under $700" — and the retailer no longer owns the first interaction. That could change the value of advertising, loyalty programs and being the default starting point for search. Mastercard is already building infrastructure to help merchants make products available to agents while keeping control of pricing, fulfillment and transactions.
What can't AI shop away?
None of this erases traditional moats — if anything, agentic AI makes it easier to tell which ones were real. Patented products cannot be compared away. A competitor cannot instantly create another well-located piece of infrastructure, another ski mountain, or another casino license in a market where licenses are limited. AI may find a cheaper vacation, but it cannot recreate the exact cruise or resort a customer wants.
For decades, some companies profited not because they offered the best product, but because consumers were imperfect shoppers — short on time, information and patience. Agents will not eliminate those imperfections overnight; consumers still must trust software with money and personal data, and companies will keep limiting what outsiders can do.
But as agents take over more of the shopping, switching and canceling, every management team now faces one question: how much of your moat comes from being genuinely better — and how much comes from customers not bothering to leave?
via fortune.com/feed/ (Source)
More from Olivia Hart
Show full bio
News editor covering regulation, policy and their impact on corporate strategy and chief-executive leadership at CEO Frontier.
456 articles
Adjacent sheets
- TikTok CEO and Microsoft Executive Warn AI Hype Outruns Reality
- Brands' Next Customer Is an AI Agent—and It Ignores Brand Size
- Only 2% of 2,053 Organizations Can Authorize Customer AI Agents
- UBS Warns Markets Underappreciate AI Agents Breaking the Human-Eyeball Internet
- AI Now Decides Which Brands Shoppers Consider — and Nike Is Losing