An AI sales agent calls a business. A real receptionist answers. The AI asks how it can help her, and the lead is gone.
It’s a real, unedited recording, and it’s the second AI call we’ve shared this year. (You can watch the first one here.)
What happens on the call
In the recording, the AI agent reaches a human receptionist, confirms it’s AI, then asks how it can help her, as if she had called it. When she points out that it called her, it thanks her for “pointing that out” and moves into a sales pitch. She declines and ends the call.
Cold calls fail all the time. This one failed because a bot didn’t seem to know why it was calling, and it cost the business a potential customer.
Why this matters for your business
It’s an awkward call to listen to. And a business paid for it. Somewhere, a company switched on an AI agent, and that’s how it’s introduced itself to a customer.
The receptionist won’t remember the pitch. She’ll remember the company that sent a bot that had no idea why it was calling. That business will never find out. Lost leads don’t show up in a report. They just don’t call back.
The same is true of AI answering your phones. Every call is a first impression, whether it’s a new lead, a regular customer, or someone with a problem they need fixed today. Get it wrong, and the person on the other end of the phone takes their business somewhere else.
But get it right, and that same call can be the start of a lasting customer relationship. A real person listens, asks the right questions, and makes the caller feel valued from the first hello. That’s what people remember.
What the research says
We surveyed 6,000 adults across the US, UK, and Canada with OnePoll. Here’s how people feel when AI picks up:
- 55% are less likely to become a customer if a business uses AI instead of people to answer calls.
- 31% would hang up if they were connected to AI when calling a business, up from 29% last year.
- 59% find AI agents frustrating when they’re calling customer service, up from 54%.
- 34% say AI customer support rarely or never resolves their problem.
- 55% are more likely to leave a review after a positive interaction with a person, compared with just 8% after AI.
The upside is just as clear. When a real person answers, customers stay, come back, and tell others. See all 45 statistics from the research.
When AI gets it wrong, someone pays
We asked real people about their experiences of AI customer service. A few stood out as the business ended up paying for them.
- Marcus in the UK needed to get in touch with his financial services provider to make a transfer. An AI assistant wouldn’t put him through to a person, and it took between 30 minutes and an hour to reach someone. By then, he couldn’t get to his own money in time to make a purchase that day. After a formal complaint, he was offered £30 in compensation. “But it cost them far more than that in my trust and loyalty,” he said.
- Fequila works in franchise sales in the US. She spent almost eight minutes stuck in an automated loop with an energy supplier, asking again and again for a person before the call cut off. The company then credited her account $20. “Not nearly as much as it cost me to miss out on work due to the issue,” she said.
- Rebecca, a mobile phone customer in Canada, was sent in circles by an AI assistant and told to wait up to 72 hours for a callback. When a person finally called three days later, her issue was fixed in three minutes. She switched providers.
The compensation in each case was small. The cost in trust was much higher, and in Rebecca’s case, the business lost a customer for good. In every story, a real person could have fixed it in minutes.
Where AI fits, and where it doesn’t
AI has a real place in business. It can speed up admin and free your team up for the work that matters most. But the phone is different. When someone calls you, or picks up when you call them, they’re deciding whether to trust you. That’s the moment that wins or loses a customer. That’s why we’ve pledged to keep every customer conversation human. When your customers call, a real person answers, listens, and knows what to do next.
If you’re using AI on your phones, it’s worth asking:
- Can it explain why it’s calling in the first sentence?
- Can callers reach a real person quickly when they ask for one?
- Are you listening back to calls, or just reading the reports?
- Would you be happy if a prospect shared the recording?
For more on getting the balance right, watch our webinar on how to resist the AI customer experience trade-off.
Make your callers day
That receptionist’s morning was interrupted by a bot that didn’t know why it called. Your callers deserve better, and so do you.
When a real person answers, your customers feel heard, and you get on with the work in front of you, knowing every call is handled. That clear, capable morning feeling lasts all day. See how we keep customer service human.


