Let me start by conceding the strongest version of the case against me.

Industry surveys — including a widely covered HubSpot/SurveyMonkey study — have found that roughly half of consumers actively dislike AI in customer-service interactions. Business press headlines have been brutal: CNBC ran "I hate customer chatbots," Forbes told small businesses their customers hate their AI chatbot. Spend ten minutes in any small-business forum and you'll find people saying they hang up the moment they hear a synthetic voice.

I sell an AI receptionist. If the research says customers hate them, either I'm selling something that damages my clients — or the research is measuring something else. I think it's the second, and I'd like to make that case carefully, because it's the whole question.

What the surveys are actually measuring

Almost all of the AI-backlash data is about deflection — enterprise chatbots and phone trees whose designed purpose is to keep you away from a human because humans cost money. You've experienced this: the airline bot that answers everything except your question, the "did you know you can check our FAQ" loop, the IVR that makes you say "representative" four times. People are right to hate that. I hate that.

A small-business receptionist is the opposite design. Nobody at the plumbing company is trying to deflect your call — the owner desperately wants your call; they're just under a sink. The honest comparison for a two-person shop was never "AI versus a friendly human at a desk." That desk doesn't exist. The comparison is AI versus a ring-out at 7 p.m. — and against voicemail, an answered call wins with almost everyone, because roughly two-thirds of callers won't leave a voicemail at all. They just dial the next result.

Where the hatred is genuinely earned

The complaints cluster into three configurations, and all three are choices, not fate:

The fake human. A bot with a human name, fake typing sounds, sometimes literal recorded office chatter. Callers figure it out mid-sentence and never forgive the business — not because a robot answered, but because the business tried to trick them. Operators who switched to announcing the AI up front report the complaints essentially stopping.

The trap. You ask for a person and it argues. You rephrase and it loops. The rage isn't about the robot — it's about being held hostage by one.

The message-taker. It chats pleasantly for ninety seconds and then delivers... "I'll have someone call you back." Voicemail with extra steps. Callers know it, and they're right.

Configure the opposite of all three — announce the AI in the first sentence, transfer instantly to a human on request, and actually book the job instead of taking a message — and you've removed nearly everything the surveys are angry about. Those three rules are how mine runs, on my own line, and they're non-negotiable in every build I do.

Who genuinely shouldn't use one

An honest answer needs this section. Skip the AI receptionist if: your calls are mostly emotional or high-stakes conversations that need human judgment from the first word (grief, crisis, contested family matters); your entire brand premium is the human touch and your customers pay for it; or you already answer essentially every call — in that case you don't have the problem this solves, and I'd tell you so on the phone.

For everyone else, the question isn't whether some customers would prefer a human. Of course they would. It's whether the customers you're currently sending to voicemail would prefer an honest robot that books their appointment in ninety seconds. The evidence — and my own call log — says yes.

Run the experiment on yourself

You have opinions about AI on the phone. Test them: call (912) 254-4358. It will tell you it's an AI in its first breath, and then it will try to be genuinely useful. Notice what you feel — annoyance, surprise, indifference — because that's the exact experiment your callers would run. Then decide. The full details live on the AI receptionist page.