Key takeaways
- The best dental uses are repetitive, non-clinical calls: NHS or private status, new-patient questions, opening hours, cancellations and after-hours routing.
- Pain, swelling, bleeding after treatment and anything clinical must go to a person or to the urgent dental route, never be assessed by AI.
- What the system says about NHS and private care must be accurate and match the practice's own published information.
- Missed calls are lost patients: many people who cannot get urgent care simply give up.
Why dental front desks are under pressure
Dental phone lines carry a heavy load of questions that have the same answer every time: are you taking NHS patients, what are your hours, how much is a check-up, can I move my appointment? At the same time, access is hard for many people. Among adults without a dentist who had an urgent dental need, only 16.0% obtained an urgent NHS appointment and 37.4% tried but were unsuccessful[1]. When an urgent appointment could not be obtained, taking no further action was the most common response, reported by 69.0% of those without a dentist[2].
The mix of NHS and private work also makes calls more complex. In 2026, 38% of dentists worked fully or mainly in NHS care, 37% fully or mainly in private care and 23% in a mixed model[3]. Callers often do not know which kind of practice they have reached, and reception spends time explaining it.
Realistic uses today
| Call type | What AI can do | When to hand to staff |
|---|---|---|
| Are you taking NHS patients? | Give the practice's current, approved answer and what to do next | Caller has urgent pain or a complex situation |
| New private patient enquiry | Explain how to register, the first appointment and where published fees are listed | Caller wants advice on a specific treatment or their suitability |
| Opening hours, parking, access | Answer from approved wording | Caller needs help getting into the building |
| Cancel or move an appointment | Take the request for staff to action, or confirm a cancellation if integrated and assured | Treatment-in-progress appointments or lab work involved |
| After-hours call with dental pain | Give the agreed urgent dental route and NHS 111 wording | Never assess the pain; always give the route |
| Recall reminder questions | Explain how to book and what the appointment involves | Anxious patients or children with additional needs |
Each of these frees a few minutes of a receptionist's or dental nurse's time. Across a busy week, that adds up to hours spent on patients in the chair or on the calls that need a person.
Where staff must stay in charge
- Any description of pain, swelling, trauma, bleeding after an extraction or problems after treatment. The system should give the agreed urgent route and pass the call to staff if the practice is open.
- Clinical questions about treatment options, suitability or aftercare.
- Treatment plans and estimates for an individual patient. NHS rules say the dental professional should tell the patient how much treatment will cost before starting[4]; that conversation belongs to the practice team.
- Complaints, payment disputes and anything that needs empathy and judgement.
- Safeguarding concerns, or a caller who seems vulnerable or in distress.
- Anxious patients, who often need a human voice to book at all.
The legal boundary matters too. The MHRA says many software and AI products used in health and social care are regulated as medical devices[5]. A system that assesses how serious a toothache sounds is doing something very different from one that reads out the urgent dental route.
Getting the words right under GDC rules
An AI receptionist speaks for the practice, so what it says is part of the practice's publicity. GDC guidance says dental practices must make clear in adverts and publicity whether the practice is NHS, mixed or wholly private[6]. If the system tells a caller 'we are taking NHS patients' when the list closed last month, that is misleading, and the practice is responsible.
Write every answer yourself
Do not let the system improvise answers about NHS availability, fees or treatments. Approve each answer in writing.
Link answers to your website
Where you publish fees and NHS or private status on your website, make the AI answers match exactly, and update both together.
Avoid treatment promises
Never let the system suggest a treatment will suit a caller. Use wording such as 'the dentist will discuss options at your consultation'.
Date the content
Keep a log of when each answer was last reviewed, and review at least monthly and whenever your NHS list status changes.
Cosmetic and private treatment enquiries
Private practices receive many calls about whitening, aligners, implants and facial aesthetics. These are tempting to automate because they are frequent and similar, but they carry advertising rules. GDC guidance says that when treatment is promoted on social media or deal sites, it must be made clear that the treatment may not suit every patient and depends on a satisfactory assessment[6]. Apply the same caution on the phone.
Example approved answer: 'We offer teeth whitening after a consultation with one of our dentists, who will check whether it is suitable for you. Our current consultation fee is listed on our website. Would you like me to pass your details to the team so they can book a consultation?'
Never let an automated system quote individual treatment prices beyond your published fee list, discuss prescription-only treatments such as anti-wrinkle injections in promotional terms, or suggest results.
Example: covering lunch and evenings
Example: a fictional mixed practice, Seaview Dental, had two receptionists and lost many calls at lunchtime and after 5.30pm. It piloted an AI receptionist for those hours only. The scope covered NHS list status, private registration, opening hours, cancellations taken as messages for the morning, and the urgent dental route for callers in pain.
The practice manager wrote every answer, the principal dentist approved the urgent wording, and a nurse reviewed transcripts twice a week. After a month, most lunchtime and evening calls were information questions that the system answered fully. Cancellation messages were actioned each morning, and several freed slots were filled from the short-notice list. Two callers described swelling after an extraction; both were given the urgent route and flagged for a call from the dentist the next morning, and the practice changed its wording so such callers were also told what to do if symptoms worsened overnight.
Data protection basics
Callers will mention their teeth, their medicines and sometimes their medical history, even when the system is designed not to ask. Treat every recording and transcript as health data. Before going live, complete a data protection impact assessment, agree a data processing agreement with the supplier that covers retention and deletion, update your privacy notice and say in the greeting that calls are recorded and an automated assistant may answer. Limit access to transcripts to the staff who need them, and set a short retention period for routine calls.
Choosing a product
Look for a system that keeps strictly to approved wording, hands off clearly and records every call for review. Ava, for instance, answers practice information and routes callers using the practice's own words, signposts 999 and NHS 111, and passes patient matters to the team without taking health details. Whatever you choose, check the points below before signing.
- Can we see and edit every answer it gives?
- What exactly happens when a caller mentions pain, swelling or bleeding?
- How does a caller reach a person when the practice is open?
- Where are recordings and transcripts stored, for how long, and who can access them?
- Does it integrate with our practice management system, and is that integration live or planned?
- What data protection and clinical safety documents can you provide?
- Which other UK dental practices use it, for which call types?
Measuring whether it helps
- Calls answered in the covered hours, compared with missed calls before the pilot.
- Share of calls fully answered without a hand-off, by call type.
- Messages and cancellations actioned by staff within the stated time, and slots refilled.
- Every call involving pain or post-treatment problems, reviewed by a dentist.
- New-patient enquiries that led to a booked appointment within a week.
- Complaints and comments about the automated assistant.
Review after six weeks with the whole team and decide whether to continue, change the scope or stop.
Patient trust
Public satisfaction with NHS dentistry is low: only 22% of people were satisfied in 2025, while 54% were dissatisfied[7]. Patients who have struggled to find a dentist may be wary of an automated voice. Be open: say in the greeting that an assistant may answer, explain how to reach a person, and make sure the answers are accurate and kind. An AI receptionist that gives a clear, honest answer at 7pm is better than an unanswered phone; one that gives a wrong answer about NHS availability does real harm.
For the GP view of the same technology, see AI phone answering for GP surgeries and safety rules for AI handling patient calls. Practices in Wales should also read bilingual phone messages.
Frequently asked questions
Can an AI receptionist book dental appointments?
Some systems can if integrated with the practice management system. Many practices start by having AI take booking and cancellation requests for staff to action, and add direct booking only once integration is proven.
Is it safe for AI to answer calls about toothache?
Only if it gives the practice's agreed urgent route and does not assess the problem. Pain, swelling and bleeding should always lead to a person or the urgent dental service.
Does an AI receptionist count as advertising under GDC rules?
What it says on the practice's behalf is part of the practice's publicity, so it must be accurate, including whether the practice is NHS, mixed or private.
References
Office for National Statistics (2026). Experiences of NHS healthcare services in England: September 2026 (opens in a new tab) ↑
Office for National Statistics (2026). Experiences of NHS healthcare services in England: September 2026 (opens in a new tab) ↑
Care Quality Commission (2026). State of Care 2025/26: dental (opens in a new tab) ↑
NHS (nhs.uk) (2026). Understanding NHS dental charges (opens in a new tab) ↑
GOV.UK (Medicines and Healthcare products Regulatory Agency) (2025). Software and artificial intelligence (AI) as a medical device (opens in a new tab) ↑
General Dental Council (2026). Guidance on advertising (opens in a new tab) ↑
Nuffield Trust and The King's Fund (2025). Public satisfaction with the NHS and social care in 2025: results from the British Social Attitudes survey (opens in a new tab) ↑