Would Your Customers Rather Talk to an AI? What the Research Says
Dallon Robinette
·
·
6 min read
Customers are skeptical of AI in customer service until they know they can reach a person. Here's what the research says, and how to set up an AI receptionist customers trust.
If you've ever called a business, heard a cheerful voice say "I'm a virtual assistant, how can I help?" and immediately started saying "representative" over and over, you're in good company. Most of us have done it. Some of us have done it louder each time, as if the problem were volume.
So it's fair that owners hesitate before putting an AI receptionist or follow-up agent in front of their own customers. The question we hear most often is some version of: "Won't my customers hate it?"
That's a question research can actually help with. Several large surveys from the past two years asked customers how they feel about AI in customer service, and when you read them side by side, a pretty clear picture shows up. Customers are skeptical of AI in general. They're much more open to it under a few specific conditions. And those conditions happen to be things a small business can control.
Here's an honest look at what the data says, and how to set up an AI agent your customers will actually trust.
The headline numbers look rough
Let's start with the number that gets quoted the most. In a Gartner survey of 5,728 customers, 64% said they'd prefer companies didn't use AI for customer service at all. Fifty-three percent said they'd consider switching to a competitor if they found out a company was going to use it.
More recent data points the same way. Verint's State of Customer Experience 2026 report, based on 5,000 U.S. consumers surveyed early this year, found that 61% prefer a human agent and 39% prefer AI. Verint also says preference for humans went up from the year before.
If you stopped reading there, you'd conclude that customers don't want AI anywhere near them. To be fair, a lot of headlines did stop there.
Then you read the reasons
The Gartner survey also asked what worries customers most about AI in customer service. The top concern was that it would be harder to reach a person. Wrong answers came in third.
In other words, the biggest fear is getting stuck. Gartner's own researcher put it well: "Many customers fear that GenAI will simply become another obstacle between them and an agent."
That points to a more useful question: what happens when customers know they can still reach a person?
What happens when there's a way out
This is where the research gets encouraging. In June, Five9 surveyed 3,000 consumers and found that 4 in 5 are willing to use AI-powered customer service when they know there's a path to a live person.
My favorite stat on this comes from the same survey. When customers were offered a clear way to reach a human, 55% said they trust the AI. When no human option was available, trust dropped to about a quarter. The AI is identical in both cases. The only thing that changed is whether the customer can get out.
Verint's numbers tell a similar story from a different angle. Among the customers who said they prefer a human, 69% said they'd happily use automated self-service if it could fully resolve their issue. And when Verint asked what customers care about most, quick resolution came first, solving the problem without needing a person came second, and easy access to a person came third. Customers want the fast answer and the escape hatch, both at once.
Customers draw the line at stakes
The other pattern in the research is that comfort with AI depends a lot on what the call is about.
A July 2026 Censuswide survey of 2,003 U.S. consumers, run for Answering Service Care, asked people to sort different kinds of calls into "AI is fine" and "I want a human." When a call carried real stakes, 67% wanted a person. For 911 calls, 57% wanted a human, and suspected fraud and debt collection each had 55% wanting one. Even for routine calls, like store hours or tracking a delivery, only about 30% said they'd actively prefer AI, so routine calls are where AI is tolerated more than requested.
For a plumbing company, a dental office or a property manager, that maps out pretty neatly. Booking an appointment, confirming a time, checking on a job's status or taking a message after hours are low-stakes and easy to get right. A billing dispute, a burst pipe at 2 a.m., or a customer who's already upset carries real stakes, and that customer wants a person who can make a decision.
And customers want to be told
The last piece is disclosure. Gartner research cited in the same CX Dive report found that about three-quarters of consumers want to know when they're interacting with content produced by generative AI. A 2025 survey by Emplifi of nearly 900 social media users found 83% want brands to disclose when they use AI, and half said a label would make them trust the brand more.
That last part is worth sitting with. A lot of businesses worry that saying "this is an AI" will turn customers off. The research suggests the opposite risk is bigger: customers who figure it out on their own tend to feel tricked, and nobody calls back a business that made them feel tricked.
A few caveats before we draw conclusions
Most of these surveys come from companies that sell customer service technology (Verint, Five9 and Answering Service Care all have products in this space), and Gartner's big survey was fielded in late 2023, which is a long time ago in AI years. Sample sizes are solid, but none of them are about small businesses specifically. So I'd treat the exact percentages as directional. What gives me confidence is that the pattern is the same across all of them, regardless of who ran the survey or what they sell.
What this means if you're setting up an AI receptionist
Here's how we'd translate the research into setup decisions for an AI receptionist, an after-hours line or a follow-up agent. This part is our view, built on the data above.
Let the AI handle it:
Booking and rescheduling appointments
Confirming times and sending reminders
Job or order status updates
Hours, service area and common questions
Taking a detailed message after hours
Route it to a person:
Billing disputes and refund requests
Complaints, or any caller who sounds upset
Emergencies and urgent repairs
Anything involving fraud, payments or account changes
Quotes that need judgment or a site visit
And a short checklist for making customers comfortable with it:
Say it's AI in the first sentence. Something simple and friendly, like "Hi, this is the virtual assistant for Smith Plumbing." It costs you nothing and it earns trust.
Offer a person, out loud, early. "You can ask for a team member at any time." Then make sure that actually works, every time.
Pass along the details. When a call moves to a person, they should see what the customer already said. Repeating yourself was one of the top four frustrations in Verint's data.
Hand off anything with stakes. Money, complaints, emergencies and anything that sounds off go to a human, no matter how well the AI could probably handle it.
Make after-hours promises you can keep. If the AI says someone will call back by 9 a.m., someone needs to call back by 9 a.m. That promise is where the trust either builds or breaks.
Have someone review a sample of calls each week. John wrote about this yesterday in Every Small Business Is About to Have Agent Bosses. Your best front-office person is usually the right one to own it.
The short answer
Would your customers rather talk to an AI? Most of them would rather talk to whoever solves their problem fastest, and they want to know a human is there if things go sideways. Set it up so the AI handles the easy, routine stuff, tells people what it is, and gets out of the way when the stakes go up, and the research says most customers will be just fine with it. Some might even stop shouting "representative."
If you're thinking about an AI receptionist and want a second set of eyes on the setup, we're happy to help.
Sources
Gartner, "Gartner Survey Finds 64% of Customers Would Prefer That Companies Didn't Use AI for Customer Service," July 9, 2024 (survey of 5,728 customers, December 2023). https://www.gartner.com/en/newsroom/press-releases/2024-07-09-gartner-survey-finds-64-percent-of-customers-would-prefer-that-companies-didnt-use-ai-for-customer-service
Josh Ballard, "AI vs. Human Customer Service: The 2026 Paradox," Verint, June 17, 2026 (State of Customer Experience 2026, 5,000 U.S. consumers). https://www.verint.com/blog/ai-vs-human-customer-service-paradox/
Bryan Wassel, "A clear path to a live agent boosts trust in AI customer service," CX Dive, July 22, 2026 (Five9 survey of 3,000 consumers, June 2026). https://www.customerexperiencedive.com/news/clear-path-to-live-agent-boosts-trust-ai-customer-service/825835/
Answering Service Care, "Should This Call Be AI or Human? Americans Draw the Line," 2026 (Censuswide surveys of 2,003 U.S. consumers and 252 business owners, July 10 to 13, 2026). https://answeringservicecare.com/insights/should-this-call-be-ai-or-human/
Customer Experience Magazine, "71% of Consumers Are Fine with AI Support If It's Upfront," July 24, 2025 (Emplifi, "AI in Social Media in 2025: What Consumers Want," nearly 900 frequent social media users). https://cxm.world/digital-experience/71-of-consumers-are-fine-with-ai-support-if-its-upfront/
Ready to make AI work for you?
Book an AI readiness evaluation. If there’s nothing worth automating, we’ll tell you.
