AI is making customers expect more and tolerate less | MarTech

AI is raising customer expectations faster than it is improving customer experience.

Nearly six in 10 consumers say knowing a company uses AI makes them expect faster responses, according to Alchemer’s “2026 AI & Customer Expectations Study.” At the same time, 89% say AI has hurt their experience with a company in at least one way.

That puts marketers and CX teams in a tough spot. The more they use AI to make things faster and easier behind the scenes, the more customers expect the experience to get faster and easier, too.

Forty-five percent of consumers said advances in AI have raised their expectations for companies overall. Nearly 30% expect feedback to be acknowledged within 24 hours, and 16% expect acknowledgment immediately.

And you better believe they still want a person available when AI isn’t fixing the problem.

Alchemer surveyed 2,009 U.S. consumers about AI’s role in customer experience. The biggest complaint wasn’t privacy, inaccurate information, or irrelevant recommendations. The winner, at 44%, was the difficulty of reaching a person when needed. That’s quite the problem, given that one goal of AI-powered CX is to lower costs via reduced human involvement.

Faster responses aren’t enough

Consumers clearly expect AI to speed up fixing problems: 38% want faster issue resolution, making it the most commonly requested improvement, while 34% want faster responses.

However, speed isn’t everything. One-fifth said AI has made issue resolution take longer, while 30% have had to repeat information they already provided.

That puts pressure on how teams measure automation.

First-response time and containment rate can make an AI interaction look successful, even when the customer returns through another channel about the same unresolved issue. A better view requires connecting those interactions and tracking repeat contacts, reopen rates, and time to resolution.

Customers: Understanding beats personalization

Consumers are really over AI that doesn’t understand what they want.

Thirty-seven percent got an AI response that didn’t understand their question or need. Twenty-six percent got wrong information, and 19% received irrelevant recommendations or responses.

It’s no surprise that 29% want AI to better understand their needs, while 24% want more relevant or personalized experiences.

A system can use plenty of customer data and still produce a bad interaction if it misunderstands the immediate intent. AI is renowned for being clueless about nuance and context, and adding more profile attributes won’t fix that.

Any AI for customer-facing interactions has to be tested to see if it understands what the customer is trying to accomplish, not just whether its answer is correct. Confidence thresholds and escalation rules become especially important when the system is unsure.

The most experienced AI users are harder to impress

The consumers most familiar with AI are its toughest critics.

Daily AI users are more likely than infrequent users to say when AI experiences feel impersonal, at 40% versus 22%. They are also more concerned about AI jumping to unsupported conclusions.

They are also more willing to pay for human access. Twenty-seven percent of daily AI users said they would pay extra for guaranteed human support, compared with 10% of infrequent users. Overall, 43% of all consumers would pay more for a product or service that guarantees access to human support.

Customers willing to pay for what they used to get for free is good news for the bottom line, but it sure doesn’t feel like a win. So, “Yay?”

The full report is available here. (No registration required)