Happy holidays to everyone except AI agents, apparently.
As AI takes over more of the customer experience landscape, replacing long hold times with immediate responses, customers are beginning to expect instant solutions across the board.
But, although customers are growing more impatient, AI isn’t getting all that much better at meeting their demands.
In fact, 57% of customers now refuse to wait longer than 10 minutes for service (compared to 50% in 2024), according a new report released by agentic AI customer support platform Forethought earlier this week.
Meanwhile, 20% get frustrated after just five minutes, up 6% from last year.
So you can imagine how short their fuses are during the chaos of holiday shopping season.
Losing Patience
In both 2024 and 2025, the top complaint among customers, according to Forethought, is that automated systems don’t know when to pass them off to a human representative.
Outdoor gear brand Cotopaxi came up with a “fallback” setting so customers always have the option of talking to a human on the CX team, said Bron Rasmussen, Cotopaxi’s CX Operations Coordinator, who was interviewed for the report.
More brands may want to consider doing the same as expectations rise. It’s not that customer service has gotten noticeably slower, according to Forethought’s report, but rather that “consumers have become markedly less patient.”
It’s not all bad news, though. Foresight actually found that, surprisingly, 72% of consumers say they’re willing to be more patient around the holidays, because they understand that it’s a time of peak demand.
Nevertheless, end-of-year holiday shopping is still a crucial time for marketers to ensure that their operations are running smoothly — otherwise, they may lose customers.
Bot Buddies vs. Matter-of-fact Machines
For brands and consumers alike, agentic AI seems like the most obvious solution.
But even when humans want to engage with AI, not everyone has the same perspective on what a successful interaction should look like.
Younger generations prefer a more human-like interaction when engaging with AI, with 43% of Gen Zers consistently saying “please” and “thank you” to chatbots, compared to only one quarter of Boomers. Boomers are also nearly twice as likely as Gen Zers to get annoyed by AI’s use of personal data for customized interactions, finding it intrusive.
For Gen Z, what breaks the experience is when their robots sound too, well, robotic.
Marketers can cater to the individual preferences of consumers by developing what Al Martin, Forethought’s VP of Customer Experience, called “adaptive and flexible” AI that shifts its communication style accordingly.
Models that are optimized for customer service can better process “the nuance of a conversation,” Martin said, “and detect whether the customer wants a more in-depth interaction, or just wants to get to the point.”
Meeting in the Middle
An AI chatbot that understands nuance sounds perfect in theory — until you start to wonder about the privacy implications.
Forethought’s solution is to train models on each individual customer’s own data, rather than pooling all customer data into a single source, and to redact any stored information to ensure that everything remains “siloed and encrypted,” said Martin.
And, of course, there will almost always continue to be problems that require human expertise to solve, he added. But if AI can perform an “initial triage,” Martin said, the human support team will have more time to spend actually helping customers instead of first having to decode their initial request.
The debate over human versus AI customer support isn’t about choosing one or the other, said Martin. “The companies that win will take a hybrid approach,” he said.
Joanna Gerber is Associate Editor at AdExchanger.
