Industries · Dental
AI for UK dental practices: the admin, not the dentistry
Every practice gets sold something for no-shows. Before any of it, there is a randomised-trial finding that costs nothing to implement and beats most of what is on the market — so it goes first on this page, ahead of anything we sell.
- Dental
- Dental practices
- Appointment reminders
- Did not attend
- Business process automation
- Recalls
The short answer
Noxia builds admin systems for UK dental practices — recall lists, reminder sequences, document and form chasing, and first drafts of routine correspondence — each with a record of what was sent and when. We do not touch clinical decisions, triage, or anything a clinician must own. Before any of it we point practices at the evidence: in a randomised controlled trial of 10,111 hospital outpatients, stating the appointment’s specific cost in the reminder cut did-not-attend rates from 11.1% to 8.4%.
On this page · 6 sections
This page starts with the thing that makes us least money, because it is the thing with the best evidence behind it.
Two randomised controlled trials, published in PLOS ONE, tested reminder wording on hospital outpatients at Barts Health NHS Trust. In the first, with 10,111 patients, adding the specific cost of the appointment to the reminder message — what the slot costs if you do not come — reduced did-not-attend rates from 11.1% to 8.4%. In the second, with 9,848, a specific cost beat a general one, 8.2% against 9.9%. No app, no AI, no deposit scheme. A sentence.
Those were hospital appointments, not dental ones. We cite the trials for the mechanism, and a practice should expect its own number to differ — which is why the change is worth measuring in your own diary rather than taking from us.
If you do nothing else after reading this page, change the wording of your reminder. It costs an afternoon and it is better supported than anything we could sell you. The full note is here.
Three figures we will not quote at you
Dental is heavily marketed to, and three of the numbers in that marketing did not survive our checking:
| The claim | What we found |
|---|---|
| 1 in 7 appointments is a no-show | Struck as a statistic. Traces to one NHS dentist describing his own practice over one year, in a trade interview. No sample, no method, not a BDA study. Fine as an anecdote, named as one. |
| £39 per missed appointment | Struck. No source found at all. The nearest real figure is £56,000 across one practice for one year — again, that same interview, again that one practice. |
| Deposits cut no-shows 55–65% | Struck, and replaced. No source for the deposit claim. The randomised trials above test something cheaper and report it properly. |
We are telling you this on our own sales page because a supplier who quotes you an unsourced number to open a conversation will quote you an unsourced number to close one. Every figure we do use is in the public register, with its source, its date and how much weight it carries. The four ways a figure fails is how we sorted these.
What we automate
| The job | What the system does | What stays human |
|---|---|---|
| Recalls | Builds the list from your practice management system on a schedule, in the right order, without anyone exporting a spreadsheet. Sends, logs, and flags the ones that bounce. | Who gets called instead of messaged, and any patient a clinician has flagged. |
| Reminders | Sends the sequence, with the wording the evidence supports, and records what went to whom and when. | The wording itself, signed off once by the practice. |
| Form and document chasing | Medical histories, consent forms, referral paperwork. Chases on a schedule until it arrives, then stops, and tells you which patients are still outstanding before the list is finalised. | Any clinical follow-up the missing document implies. |
| First drafts | Routine correspondence — referral letters, treatment-plan summaries — assembled from the record, marked as a draft. | Every word that goes out. A clinician edits and owns it. |
What we will not build for a practice
Anything that touches triage, diagnosis, or clinical priority. Anything that decides which patient is seen first. Anything that answers a clinical question from a patient without a clinician reading it. These are not technical limits; they are the line, and it does not move for a discount.
The record of what the system did — and what it declined to do — is written as it happens rather than reconstructed afterwards. That is the audit layer, and it is the same one our regulated-firm clients get.
An honest word about size
Our own research database holds 28,583 UK organisations and 7 of them are dental practices, against roughly 25,500 in financial services. We lead with financial services because that is where our evidence and our experience are, and it would be dishonest to present this page as a specialism.
What we do have is the admin pattern, which is the same across a diary-run business, and the discipline about figures, which is rarer than the software. If you want the arithmetic before the conversation, the missed-call calculator takes six numbers you already have and can tell you to buy nothing, and the triage tool will talk you out of most of it.
Questions people actually ask
What is the most effective way to reduce dental no-shows?
On the best available evidence, changing the reminder wording. In a randomised controlled trial published in PLOS ONE, among 10,111 hospital outpatients, stating the appointment’s specific cost in the reminder reduced did-not-attend rates from 11.1% to 8.4%, and a second trial found a specific cost beat a general one. The trials were in hospital clinics, not dental practices, but the change costs nothing to test.
Is one in seven dental appointments really a no-show?
We could not verify it. The figure traces to a single NHS dentist describing his own practice over one year in a trade interview — no sample, no method, and not a British Dental Association study. It is a reasonable anecdote and a poor statistic, and we do not use it.
What dental admin can safely be automated?
Recall list building and sending, reminder sequences, chasing medical histories and consent forms, and first drafts of routine correspondence for a clinician to edit and own. Triage, diagnosis, clinical priority, and answering a patient’s clinical question without a clinician reading it should not be automated.
Does Noxia specialise in dental practices?
No, and we will not claim to. Our own research database holds 28,583 UK organisations of which 7 are dental practices, against roughly 25,500 in financial services. The diary-and-chasing admin pattern is the same, and that is what we bring; the sector depth is in financial services.
Sources
- Hallsworth M et al., two randomised controlled trials published in PLOS ONE, at Barts Health NHS Trust outpatient clinics: in the first, 10,111 patients, including the specific cost of the appointment in the reminder reduced did-not-attend rates from 11.1% to 8.4%; in the second, 9,848 patients, a specific cost beat a general one, 8.2% against 9.9%. doi.org ↗ — primary; peer-reviewed randomised controlled trials — in hospital outpatients rather than dental patients — and the strongest evidence on this page re-checked yearly
- Noxia figures register, FIGURES.md, rendered at /what-we-checked. “1 in 7 dental appointments a no-show” struck as a statistic — traces to one NHS dentist describing his own practice over one year in a trade interview, no sample, no method. “£39 per missed UK appointment” struck — no source found; nearest real number £56,000 across one practice for one year, from the same interview. “55–65% no-show reduction from deposits” struck and replaced by the randomised trials above. noxia.co.uk ↗ — primary and first-party: our own record of the figures we removed
- Noxia research database, Lead_Magnet/data/leadmagnet.db, counted by SQL on 7 September 2026: 28,583 UK organisations, of which 7 dental practices, 16 letting agents, 4 plumbing and heating, and approximately 25,500 in financial intermediation, insurance broking, mortgage finance and fund management. — first-party; our own database, counted rather than estimated, and quoted here against our own interest re-checked every 6 months
- The division of jobs into what we automate and what stays with a clinician is ours. It is not a regulatory standard and no dental body has endorsed it. — our opinion, labelled
Checked 24 September 2026. Next scheduled check 23 March 2027. Numbers that move — leaderboards, live indices — are re-checked every 30 days; annual datasets and rules in force every six months; dated research once a year. If something here has gone stale before we got to it, tell us and we will correct it and say what changed.
Change the reminder first. Then talk to us if there is anything left.
We would rather you implemented a free intervention that works than bought a system that might. If the recalls, the chasing and the paperwork are still eating a day a week after that, tell us which one and we will say whether it is worth building.
Start the conversation