Imaging & diagnosticsVendor-reported

Pearl's X-ray AI at Blue Court Dental: what the vendor reports

Pearl's case study says that from October 2024 to January 2025, at Blue Court Dental in Harrow, Second Opinion marked 44 early carious lesions on 145 X-rays from 35 patients, and that the dentists had found 12 of those with traditional methods. The page gives no price. The practice website does not mention Pearl.

Published 22 October 2026 · Last checked 22 October 2026 · Fact-checked against 2 sources · How we verify

Practice: Blue Court Dental
Size: 35 patients in the report
Location: Harrow, UK
Tool: Pearl Second Opinion
Cost: Not disclosed
Source: Vendor-reported

Did it work? Results

Pearl's case study says that from October 2024 to January 2025, across 145 X-rays from 35 patients, Pearl identified 44 early carious lesions.
Source
Pearl's case study says the dentists had spotted 12 of those 44 lesions with traditional methods, and that the AI caught 32 lesions that might have been missed.
Source
Pearl's case study calls those 32 lesions a 73% detection rate. It does not define that rate against a reference standard.
Source
Pearl's case study says nine of those missed lesions required immediate reassessment and restoration.
Source
Pearl's case study quotes Dr. Dixit that Pearl has sped up the diagnosis process a lot. It gives no minutes saved.
Source
Pearl's case study says sales of fluoridated mouthwashes, specialized floss, and regular hygiene maintenance appointments increased. It gives no counts.
Source

Who tried it, and why

General and cosmetic dental practice. The Pearl page reports 35 patients and 145 X-rays in this review. It does not give the size of the practice.. Harrow, UK (Pearl's case study and the practice website).

Pearl's case study says Dr. Nishan Dixit, at Blue Court Dental in Harrow, was recommending preventive treatment that patients declined because the tooth did not hurt yet. The page says patients are not trained to read X-rays, so they could not see the early decay he was describing.

What did they set up?

Pearl says Dr. Dixit first saw Second Opinion at a conference in Chicago. The page does not date that conference.
From October 2024 to January 2025, the page says he used it on 145 X-rays from 35 patients, a mix of new-patient assessments and routine recalls.
The page describes the display as color-coded X-rays with indicators on areas of concern, used while talking with the patient. It says he could show decay progression percentages. It does not give an example percentage.
Not disclosed: the price, which X-ray system the images came from, whether every film in the practice was run through the AI, or how the "traditional" read was recorded.

Tools used

Pearl Second Opinion (AI review of dental X-rays)

What did it cost?

Not disclosed. The case study gives no price.

“Patients are not trained to read X-rays.”
Dr. Nishan Dixit, Blue Court Dental, quoted in Pearl's case study

What didn't go well?

Every count is published by Pearl. The page does not show the 145 X-rays or say who marked the "traditional" findings, or whether they were marked before the AI was shown.
The page's "73% detection rate" is its label for the 32 additional lesions out of 44. 32 of 44 is about 73% by ordinary rounding. That is not a sensitivity measured against a reference standard, and the page does not provide one.
No price. The page also gives no count for the extra mouthwash, floss, or hygiene visits it says increased.
The practice website, read on 2026-10-22, does not mention Pearl. It says Blue Court Dental Centre in Harrow was established in April 2000 by Dr Nishan Dixit. Pearl's page says "25+ years" of experience. Those two descriptions are not the same wording.
The page says patients could be shown precise decay progression percentages. It never gives a percentage from a real film.

What should your office take from this?

Before you trust a detection count, ask how the comparison was made: same films, read before the AI was turned on, by whom, and against what. This case study does not say.
Use the color overlay to explain a finding. The page's account is about showing early decay, not about the AI treating anyone.
Ask for the monthly price, which X-ray systems it reads, and whether a dentist still has to confirm every mark. None of that is in this case study.
Check the practice's own site and your own films. Blue Court Dental's website does not mention Pearl.

Try it yourself

Questions to ask before you buy an AI second read for X-rays

You are helping a dentist prepare for a demo of an AI that marks findings on dental X-rays. Using only the practice details below, draft a one-page test plan. Include: (1) 10 films we will run, described by type only (bitewing, periapical, panoramic), with no patient name and no image attached; (2) the rule that a dentist records a finding list before looking at the AI, then compares; (3) what we count as a match, a miss, and a mark we reject; (4) what the AI must not do (diagnose without a dentist, recommend treatment on its own, or store images without our written terms); (5) the price, the X-ray systems it accepts, and who signs the agreement. Ask me for anything missing. Do not ask me to paste an X-ray or any patient information.

Practice details: [X-ray system] [who reads films today] [what we want marked]

Prompt written by us, not by the practice. Never paste names, dates of birth or other patient details into a general AI tool.

Sources

Catching Cavities Before Crisis: How Pearl Empowers Minimally-Invasive Dentistry at Blue Court Dental
Pearl · Undated; accessed 2026-10-22
Blue Court Dental Centre
Blue Court Dental · Site says the practice was established in April 2000; accessed 2026-10-22

How we checked this

Checked on 22 October 2026. We read all 2 sources listed above, and a separate second check re-opened every one of them before publication. Anything we could not confirm was removed.

Figures published by a vendor are reported as that vendor's claims, not as verified facts.

Spotted an error? Tell us and we will re-check it. How we verify.

Free assessment. Tell us what takes the most time in your office. We will reply with what we would automate first.Get a free assessment →

Get one real case study a week

More case studies

Front desk & phonesVendor-reported

AI receptionist for overflow and after-hours calls: Wolfe Dental's results

A two-office practice in the Portland area with one receptionist per office let Arini's AI take overflow and after-hours calls. Arini reports missed calls fell from as many as 500 to about 50 a month and $134k in booked production over eight months, but it gives no price.

Arini reports the AI answered an average of about 500 overflow and after-hours calls a month that had previously gone to voicemail.

Front desk & phonesMixed

AI receptionist audit: LA dental office's 15-month call log shows 41% task completion

A part-time staff member at an Open Dental practice near Los Angeles read every failed call from 15 months of an AI receptionist. It finished 263 of the 637 calls where someone wanted something done, never put a caller through to a person, and had trouble with Spanish and with spelling names. The author later built a competing product, so weigh the comparisons with that in mind.

Completion rate: of the 637 calls where the caller wanted something done, the AI completed 263 (41%). The 263 includes 76 calls where it only took a message and 45 information-only calls. Actual bookings, reschedules and cancellations made on the call totalled 142, about 22% of the 637 (our arithmetic from the author's table).

Front desk & phonesMixed

Dental Intel AI receptionist vs. Mango Voice Margo: what dental offices report

Some offices are trying the AI receptionists that their current phone or practice-platform vendor now sells as an add-on. One office uses Dental Intel's only for overflow and after hours, and it books new patients at night. Another tester couldn't reach the doctor in a mock emergency. A Mango Voice office says Margo gets clearer caller details than staff do, but it can't write back to Open Dental and its calls are hard to track.

Practitioner report (Dental Intel): about a month after starting, using it only for overflow and after hours, the office says the AI had booked a number of patients, including new patients after hours, and fills openings in the hygiene schedule.