For years, subscription businesses have benefited from a simple behavioral advantage: consumers are far better at signing up than they are at canceling.
People forget what they joined. They stop using services but keep paying. A $9 or $15 monthly charge can linger on a credit card statement for months, sometimes years, without attracting much attention.
Follow THE FUTURE on LinkedIn, Facebook, Instagram, X and Telegram
Meta’s Muse Targets A Deeply Embedded Habit
With the rollout of Meta Muse, an AI personal agent designed to handle tasks across everyday life, that long-standing subscription friction is suddenly under pressure.
Muse can help consumers identify and cancel recurring subscriptions. Subscription-management tools have existed for years, but Meta’s broader AI assistant could make forgotten charges far easier to find and remove.
The timing is significant. Nearly half of U.S. consumers, 44%, increased subscription spending in 2025, according to a report from Mastercard and FT Strategies. Average annual spending rose to $1,887, or about $157 a month. Bank of America data also shows subscription spending climbed 7.7% year over year in July, outpacing overall card spending, with entertainment and retail subscriptions accounting for roughly 43% of the total.
The Economics Of Forgetting Are Under Pressure
Neale Mahoney, an economics professor at Stanford University and director of the Stanford Institute for Economic Policy Research, has spent years studying how much subscription businesses benefit when consumers fail to cancel.
“We found that when people are forced to decide, they are about four times more likely to cancel,” Mahoney said, referring to his 2025 American Economic Review paper, Selling Subscriptions, co-authored with Stanford economists Liran Einav and Ben Klopack.
That finding matters in an AI-agent world.
The Stanford researchers estimated that sellers can roughly double revenue through consumer inertia — when people forget to cancel — and through cancellation friction, when ending a subscription takes too much time or effort. AI personal agents could weaken both forces, Mahoney said.
Not every subscription is equally exposed. Physical subscriptions, such as pet food, are harder to forget when deliveries keep arriving at the door. Digital services, by contrast, can quietly continue billing long after a consumer has stopped thinking about them.
The implications could extend well beyond subscriptions and into the core of financial services. Apollo chief economist Torsten Slok recently warned that AI agents could automatically move household cash into higher-yield accounts, reducing the cheap deposits banks rely on to fund loans.
Consumers Are Already Getting More Aggressive
Even before AI entered the picture, consumers were becoming more assertive about cutting recurring charges.
ScribeUp, which builds subscription-management technology into banking apps for banks, credit unions, and fintechs, says its members are now 1.8 times more likely to initiate a cancellation than they were a year ago.
“More of life is becoming a recurring bill. Those bills cost more every year, and it becomes increasingly difficult for consumers to keep track of everything they are paying for,” said Jordan Mackler, co-founder and CEO of ScribeUp.
According to the company, the median ScribeUp user now has more than 12 recurring subscription payments, while one in four has 20 or more. The share of users with at least eight recurring charges rose from 62% to 71% over the past year.
Price increases can accelerate the exit. Mackler said cancellations at an individual merchant can rise as much as 50% when prices go up.
Health and fitness subscriptions have seen the sharpest jump in cancellation activity, up 3.8 times year over year. Video streaming followed at 2.2 times, news and media at 2.1 times, and music streaming at 1.9 times, according to ScribeUp.
AI is also expanding the number of subscriptions the company can help manage. ScribeUp now tracks roughly 200,000 unique recurring billers. Before adding its newer agentic capabilities this year, it could automate cancellations for only a few hundred major subscription companies.
The average canceled subscription costs $17.39 a month, and Mackler said ScribeUp saves users more than $300 a year on average in recurring charges they had forgotten about or no longer needed.
Subscription Businesses May Need A New Retention Playbook
The subscription model is already dealing with meaningful churn. Mastercard found that average monthly churn — the percentage of subscribers who cancel or fail to renew in a given month — is 20%. More than half of U.S. subscription businesses surveyed also said at least 10% of their subscriber base was inactive, meaning customers were still paying but no longer actively using the service.
As consumers gain greater control over recurring charges, companies may need to compete harder to keep them.
Hitee Chandra Jha, principal product manager at Zendesk who specializes in product-led growth and customer retention, said companies may need to make value more visible before customers reach the cancellation screen.
For some businesses, that could mean treating cancellation less as a binary outcome and more as a transition. A streaming service might do better offering a pause option than a last-minute discount after a customer finishes a series. A fitness app could move a user who has reached a goal to a maintenance tier or offer a different way to stay engaged.
That approach appears to be gaining traction. Recurly’s 2026 State of Subscriptions report, based on 76 million unique subscribers across more than 2,200 businesses, found that use of “pause before cancel” options jumped 337%, and three out of four customers who paused eventually returned.
The Best Defense May Be Convenience, Not Friction
Meta CEO Mark Zuckerberg has described Muse as the “centerpiece” of an AI strategy that will be supported by still unproven consumer hardware, including augmented reality glasses and the Muse Charm keychain-style device. Investors have recently rewarded the company for that AI bet.
The rollout has not been without resistance. Amazon has blocked Muse from shopping on its site, saying the agent’s access violates its terms of service. Privacy concerns also remain significant whenever an AI system gains access to financial information.
Still, consumer interest appears real. According to Recurly data, 43% of consumers said they are comfortable with AI managing their subscriptions.
“Treating cancellation as a transition, not just a loss, is what differentiates a mature retention strategy from a defensive one,” Jha said. “The best defense against AI-assisted cancellation is not friction.”
Mastercard’s research points in the same direction. Among consumers surveyed, 74% said they are more likely to subscribe when cancellation is easy, while 70% said they are more likely to resubscribe. Another 34% said they would remain subscribed if they had the option to pause rather than cancel.
Rather than ending subscription growth, AI may force companies to prove value more consistently and design retention strategies that depend less on inertia.
ScribeUp has found that once users trust they can see, manage, and cancel recurring charges easily, they become more comfortable with subscriptions overall and are more willing to try new services.
“If consumers trust that they can easily see, manage, and cancel recurring charges, our data shows they are likely to expand their overall recurring spend and are more willing to try new services in the first place,” Mackler said.
Mahoney argues that a healthier market would reward businesses that do not rely on slowing customers down, but instead give them more reasons to stay.
“When people are trapped in subscriptions they don’t want or can’t get out of, market forces are limited, and companies don’t have the incentives to have a high-quality product at a low price,” he said. “They can just rely on a locked-in install base.”
Whether through AI or consumer initiative, when people stop paying for what they no longer want and redirect that money elsewhere, “that’s good for consumers, it’s good for markets, and it’s good for firms that are making products that people want,” Mahoney said. “That’s economics working the way it’s supposed to work.”
For Meta, Muse may be more than another AI product launch. It could become a test of whether artificial intelligence amplifies consumer discipline — and forces an entire subscription economy to compete on value rather than habit.









