Will Patients Be Annoyed If AI Answers?

The Short Version

This is the objection that stops most practice owners, and it deserves a real answer rather than reassurance from a vendor.

The public opinion research says three things at once, and they do not all point the same way.

1. Scheduling is the AI use case people are least concerned about. In Pew Research Center’s survey of 3,488 US adults, fielded 22 to 28 June 2026 and published 25 August 2026, 56% said it is extremely or very important to be told when AI is used to schedule their appointments. For AI making diagnoses or analysing scans, that figure was 81%. A third of respondents, 33%, said they do not feel they need to be told about AI scheduling at all.
2. General comfort with AI in healthcare is falling, not rising. An Ohio State University Wexner Medical Center survey of 1,007 adults, fielded 16 to 20 January 2026 and published 7 April 2026, found 42% of adults supported AI use in healthcare, down from 52% in 2024. Margin of error was plus or minus 3.5 points.
3. The thing people object to most is not being told. In the same Pew survey, 53% said they have not too much or no say in whether AI is used in their care, and 63% said they want more say.

Put together, the finding that matters for your phone is this: the annoyance is mostly about concealment and control, not about the automation itself, and answering the phone sits at the mild end of what people worry about.

That is good news for a practice considering this, but it comes with a condition attached, and the condition is the whole ballgame.

Why Scheduling Is Different From the Scary Cases

It is worth understanding why the numbers split the way they do, because it tells you where the real risk is.

The uses that alarm people are the ones where AI substitutes for clinical judgement. Pew found 81% wanted notification for diagnoses and scan analysis, and 80% for explaining test results. Those are decisions with consequences the patient cannot check, made about their body, by a system they did not choose.

Booking a hygiene appointment for the fourteenth is not that. The patient knows immediately whether it worked. If the appointment is wrong, they find out at the next step and it gets fixed. There is no hidden clinical consequence, and the feedback loop is immediate.

This distinction is also the line your own deployment should respect. An AI receptionist that recognises a caller is describing facial swelling and routes it to your protocol is doing scheduling and triage routing. An AI that tells a patient what their swelling probably is would be doing something the survey data says people are genuinely not comfortable with, and which they are right not to be comfortable with. We write about where that line sits in our piece on handling dental emergencies after hours.

What Actually Annoys Callers

Set the survey data aside for a moment. In the specific setting of a dental phone line, the complaints that come up are rarely “this is AI.”

Being trapped. The single most reliable way to annoy a caller is to make reaching a person impossible. This is the old IVR grievance and it long predates AI. “Press 1 for appointments” with no route to a human is the thing patients learned to hate, and an AI that cannot hand off recreates it with a friendlier voice.

Being fooled. Discovering afterwards that the warm, competent person was software produces a specific kind of irritation, because the patient feels they were managed. The Pew finding that 63% want more say is this feeling measured at scale.

Repeating themselves. Explaining the problem to the AI, getting transferred, and explaining it again to a human. This is a handoff design failure and it is entirely avoidable.

Being misunderstood on something urgent. A caller in pain who has to fight the system to be taken seriously will remember it, and will tell people.

Notice that only one of those four is inherent to AI. The other three are configuration choices, and you control all of them.

What Reduces the Annoyance

Five things, in rough order of how much they matter.

1. Say what it is, in the first sentence. The AI should introduce itself as the practice’s virtual receptionist. Not a fake human name with no qualifier, and not an evasion if asked directly. This is the single highest-leverage choice, and the Pew data is the argument for it: disclosure is what people say they want, and it is nearly free to give them. Should the AI tell callers it is AI covers the disclosure duties that now apply in some states.

2. Make the human route obvious and short. A caller who asks for a person should get one, quickly, without an argument. If your team is closed, the AI should say so plainly and say when someone will call back, rather than looping.

3. Carry context across the handoff. When a call transfers, your team should already know who is calling and why. This is what turns a transfer from an annoyance into a service.

4. Answer fast, and never queue. A large share of the goodwill here comes from something unrelated to AI: the phone gets answered immediately, at 7pm, on a Sunday, without hold music. Patients notice that more than they notice the voice.

5. Get the urgent cases right. Triage routing is where a bad deployment does actual damage. Configure it on your own protocol and test it before launch.

We have two word-for-word scripts for the disclosure and the human-request handoff in our guide to building patient trust with an AI receptionist, which is the practical companion to this page.

What We Will Not Claim

A few honest limits, because this is exactly the topic where vendors overreach.

We cannot tell you your patients will like it. The survey data above is national and general. It is not dental, it is not your town, and it is not your patient base. A practice with a older patient population in a small community may get a different reaction than a downtown practice with a young professional list.

Sentiment is moving against AI generally, not toward it. The Ohio State finding of support falling from 52% to 42% in two years is not a number we would choose to publish if we were only selling. It is real, and it means the disclosure discipline above matters more each year, not less.

Nothing here is measured head-to-head. We have not run a controlled study of patient satisfaction with our AI versus your front desk, and neither has anyone else we are aware of. Treat any vendor who claims otherwise with suspicion, and ask to see the method.

The fix for a bad reaction is to stop. If you pilot this and your patients hate it, the correct response is to turn it off, not to push through. A month-to-month arrangement with no long contract is what makes that possible, which is one reason we sell one.

How to Find Out for Your Practice

Rather than arguing about it internally, run a small test.

Start with the calls you are already losing. After hours and lunch. Nobody is currently answering those, so the comparison is not “AI versus our lovely front desk,” it is “AI versus voicemail.” That is a much easier bar and a much fairer test of whether patients object.

Call it yourself five times, as a patient would. Book a new patient appointment. Ask to move an existing one. Ask a question the AI will not know. Ask for a human. Describe an urgent problem. If any of those five annoys you, it will annoy a patient.

Ask the patients who used it. At the next visit, ask the ones who booked after hours how it went. Not a survey, just a question at the chair. You will get a clearer signal in two weeks than from any amount of speculation.

Watch for the specific complaint, not the general worry. “I did not know it was a computer” is a disclosure fix. “I could not get a person” is a handoff fix. “It booked me with the wrong hygienist” is an integration fix. Each is solvable, and treating them as one vague objection is how practices talk themselves out of something that was working.

Key Takeaways

• Pew found 56% want to be told about AI scheduling, versus 81% for AI diagnoses. Appointment booking is at the mild end of public concern.
• 53% feel they have little or no say over AI in their care and 63% want more, so concealment is the thing to avoid.
• Overall support for AI in healthcare fell from 52% in 2024 to 42% in early 2026 by Ohio State’s measure. The trend is not in AI’s favour.
• Three of the four common annoyances are configuration choices: being trapped, repeating yourself, and being fooled.
• Test on after-hours calls first, where the honest comparison is against voicemail.

Frequently Asked Questions

Will patients be annoyed if AI answers our phone? Some will, and the research suggests fewer than practice owners expect for scheduling specifically. Pew’s June 2026 survey found scheduling to be the AI use case people are least insistent on being told about, well below diagnosis or scan analysis. The larger risk is not disclosing it, since 63% of respondents said they want more say over AI use in their care.

Should the AI tell callers it is an AI? Yes. It is the cheapest thing you can do to address the concern the data actually identifies, and an AI that dodges the question when asked directly creates a worse problem than the one it avoided.

What if a patient asks for a human? They should get one promptly, with context carried across. If your office is closed, the AI should say so and commit to a callback time rather than looping the caller.

Are older patients more likely to object? We do not have dental-specific data on that and will not invent it. Pew’s 2026 survey does report differences by demographic group in how much say people feel they have, with 60% of White adults reporting minimal control compared with 40% of Hispanic adults. That is about perceived control, not about age and phone systems, so treat it as context rather than an answer.

Is it worse than voicemail? For the calls currently going to voicemail, this is the comparison that matters, and it is a low bar to clear. A caller who books an appointment at 9pm has had a better experience than one who left a message.

How would we know if it is hurting us? Ask patients at the chair, watch your new patient numbers, and read the call transcripts. If patients are hanging up early or asking for a human on most calls, you have a configuration problem worth fixing or a decision to make.

Where to Start

Test it on the calls nobody is answering today, disclose plainly, and make the human route short. If those three are true, this objection mostly takes care of itself.

The two scripts are worth stealing before you go live. You can also call our AI yourself and judge the experience as a patient would, which is a more useful data point than anything on this page.

---

Sources

• Pew Research Center, “Americans want transparency when AI is used in their healthcare,” published 25 August 2026. Survey of 3,488 US adults on the American Trends Panel, fielded 22 to 28 June 2026.
• The Ohio State University Wexner Medical Center, AI in health care survey, published 7 April 2026. Survey of 1,007 US adults, fielded 16 to 20 January 2026, margin of error plus or minus 3.5 percentage points.

Both surveys measure attitudes to AI in healthcare generally, across all settings, not to AI receptionists in dental practices specifically. No dental-specific survey of this question was available to us at the time of writing, and we have not run one. We sell an AI receptionist, so read the section above on what we will not claim.