79% of Your Customers Would Rather Talk to a Human. So Why Are You Hiding the AI?

I read a stat recently that should be pinned above every CX roadmap in the country: 79% of people say they'd strongly prefer speaking to a human over an AI agent. Not "slightly prefer." Strongly. And 84% believe a human will give them more accurate support, full stop.

It's from a December 2025 SurveyMonkey study of over 2,000 people, and it lands right in the middle of the biggest AI investment wave CX has ever seen. So, while every vendor pitch in the country tells you customers are ready for AI, customers themselves are telling researchers something rather different.

Digging deeper, I found the number that I think actually explains it: 81% believe businesses are deploying AI primarily to save money, not to improve service. Suggesting that customers aren't rejecting the technology, they’re rejecting the perceived motive.

The trust gap isn't about capability anymore

For years, the excuse for bad AI experiences was that the technology wasn't good enough yet. That excuse is mostly gone. Modern conversational AI can hold a natural conversation, understand intent, pull live account data and resolve a real issue end to end.

So why does a separate US survey find 93% of people still say they'd rather talk to a human, and half say they'd cancel a service that relied on AI alone?

Because good CX was never just about resolution. It was about whether the customer trusted the organisation on the other end of the line. And right now, most deployments are eroding that trust rather than building it. Only 54% of consumers say they even want to be told clearly when they're talking to AI. Fewer than half feel confident they could tell the difference if nobody said anything.

That's the real risk. Not that AI fails to resolve the query. That it resolves the query while quietly telling the customer: we didn't think you were worth a human.

What separates the AI deployments people trust from the ones they don't

The organisations getting this right share a common pattern, and it isn't "using less AI."

They're upfront about when AI is involved rather than disguising it as a person. They make the route to a human fast and obvious, not buried behind three menus and a guilt trip. And critically, they use AI to make the human interaction better when it happens, not to remove the option altogether.

63% of people say they don't expect AI to fully replace humans in customer service. They're not asking you to choose. They're asking you to be honest about which one they're getting, and to make the handoff painless when they want the other one.

This is precisely the design philosophy behind Zoom Virtual Agent and Zoom AI Expert Assist. ZVA isn't built to trap a customer in a bot loop and call it deflection. It's built to resolve what it can resolve cleanly, and hand off what it can't, with full context, to a human who picks up the conversation rather than restarting it. AI Expert Assist does the same thing in reverse: it makes the human agent faster and sharper in real time, so when a customer does reach a person, that person is genuinely better equipped, not just available.

That's the difference between AI that earns trust and AI that erodes it. One says: "We invested in making your experience better." The other says: "We invested in needing fewer of us."

So?

When 81% of your customers already assume your AI investment is a cost play, the deployment itself is doing PR work whether you planned for it or not. Every hidden bot, every dead-end menu, every "I'm sorry, I didn't understand that" confirms the assumption. Whereas, every transparent handoff, every AI that visibly makes a human interaction faster and more informed, swiftly disproves it.

The organisations winning the trust argument in 2026 aren't the ones using AI the least, or the most. They're the ones being straight with customers about what they're getting, and making sure the option to reach a person is never more than a sentence away.

Deploy AI to save your customer time. Don't deploy it to save yourself the conversation.

Acceleraate helps organisations deploy Zoom Virtual Agent and AI Expert Assist in a way that builds customer trust rather than eroding it, transparent, fast to escalate, and genuinely useful for both customers and agents. If your AI strategy needs a trust check, let's talk.