Pearl Logo
Skip to content
Sign In To Access All Articles
  • There are no suggestions because the search field is empty.

AI Opportunities Logic for Dashboard Metrics

 

Note: The AI Opportunities Dashboard is available for Enterprise clients only. 

How We Determine Treatment Opportunities

Pearl’s AI automatically identifies and tracks potential treatment opportunities for your patients over time. Here’s how we decide when to create, update, or close an opportunity so your dashboard stays accurate and meaningful.

Tracking Changes Over Time

When our AI detects conditions on the same tooth across multiple visits, we only keep the most relevant version:

  • If a weaker finding (like caries) is later replaced by a stronger one (like a crown), we archive the older opportunity.
    • → Example: Incipient Caries → Progressed Caries → Crown
  • This prevents duplicate or outdated opportunities from appearing.

How We Track and Resolve Opportunities

Each opportunity follows a simple path from detection to completion. Pearl automatically updates the status as your team plans and finishes treatment using CDT codes from your Practice Management Software (PMS).

Opportunity Stages

  1. Unplanned – The AI detects a condition (like decay or periapical radiolucencies) on a particular tooth and there is no matching treatment in the treatment plan.
  2. Planned – Your team treatment plans a matching CDT code (for example, D2391 for a filling). This tells us the treatment has been planned.
  3. Scheduled – The treatment is attached to an upcoming appointment in your PMS. This reflects that the opportunity has been scheduled.
  4. Completed – Once the treatment is marked complete in your PMS, the opportunity is closed and moves to completed.

Example:

  • AI finds decay → opportunity created → Unplanned
  • Treatment D2391 added → Planned
  • Appointment booked → Scheduled
  • Treatment walked out → Completed

Counting Opportunities

We count each opportunity once per patient, even if it appears across multiple appointments or x-rays.

Important: The reporting date range determines which patients' opportunities are displayed on the dashboard, based on whether they had an appointment within that period. The date range is not based on when the x-rays were taken or when the opportunity was detected, treatment planned, or scheduled. 

Example: Sam Smith’s Appointments for tooth #30

  • Jan 10: Images taken. Restoration opportunity detected on tooth #30.
  • Mar 1: Restoration opportunity on tooth #30 is unplanned, then added to the treatment plan and scheduled for Mar 15.
  • Mar 15: Restoration on tooth #30 completed.

Dashboard Reporting Date Range Examples

  • Report run on Feb 1, covering Jan 1 – Feb 1:

 → One opportunity counted (patient had Jan 10 appointment).
→ Status = Unplanned.

  • Report run on Feb 15, covering Jan 15-Feb 15: 

→ This opportunity will not be counted since the patient did not have an appointment during this time period. 

  • Report run on Apr 1, covering Jan 1 – Feb 1:

 → Same opportunity counted (patient still included because of Jan 10 appointment).
→ Status = Completed.

  • Report run on Apr 1, covering Jan 1 – Mar 31:
    → Same opportunity counted once (patient had Jan 10 appointment).
    → Status = Completed.

Types of Opportunities

Incipient Caries, Restorations (Progressed Caries), Crowns:

  • Pearl uses our pathology detection models and statistical modeling to determine likelihood of detections that should be addressed. Not every detection is an opportunity, we focus on those with high statistical probability.
  • Expanding logic to include Depth and Penetration percentage E1, E2, D1, D2, or D3.

Implant: Pearl looks for missing teeth, excluding existing implants and bridges, on PA, Bitewing, and Pano images.

Impaction: Pearl looks for conditions of soft, partial bony, and full bony impactions for patients age 16+ in Pano images. 

Endo: Pearl looks for periapical radiolucency, currently Panos, expanding to PAs. 

SRP: Pearl looks for 3+ instances of >3.0 mm BioWidth or calculus on bitewing images per quadrant.

Dashboard Definitions 

  • AI Treatment Planning Rate: The percentage of Pearl AI Opportunities that were Treatment Planned. Normal range: 15-35%.
  • AI Case Acceptance Rate: The percentage of Pearl AI Opportunities that went from Treatment Planned to Scheduled. Normal range: 30-50%.
  • AI Opportunity Completion Rate: The percentage of Total AI Opportunities that were completed and posted to the ledger. Normal range: 5-15%.
  • Total AI Opportunities: Unique opportunities for patients who visited in this time range.
  • Planned Treatments: AI opportunities that were treatment planned.
  • Unplanned Opportunities: AI opportunities without treatment plans.
    • These may include opportunities that haven’t been reviewed, were reviewed and were not treatable yet, or were referred to a specialist. 
  • Planned, Not Scheduled: AI opportunities that have been treatment planned, but have no future appointment date.
  • Scheduled Treatments: AI opportunities that have a future appointment date.
  • Scheduled, Not Completed: AI opportunities that have a future appointment date, but are not completed.
  • Completed Treatments: AI opportunities completed and posted to ledger.
  • Archived Opportunities: AI opportunities that are no longer actionable.
    • Reasons: patient leaves the practice, patient rejects the treatment, alternate treatment negates / resolves this opportunity.