Dentem.ai — Implementation Playbook (DSO)

Dentem.ai Implementation Playbook for DSOs

Diagnostic Imaging AI Deployment Guide

---

1. Executive Summary

What Dentem.ai Does

Dentem.ai is a diagnostic imaging AI platform that analyzes dental radiographs in real-time to detect pathology, caries, bone loss, and other clinical findings with consistent accuracy. The system overlays AI-generated annotations on X-rays, providing chairside decision support that enhances diagnostic confidence and standardizes clinical interpretation across providers.

Why DSOs Specifically Benefit from Diagnostic Imaging AI

Diagnostic imaging AI delivers outsized value at scale for three reasons:

1. Standardization of Clinical Quality: Across 15–50 locations with varying provider experience levels, AI creates a consistent diagnostic floor—every patient receives the same caliber of radiograph analysis regardless of which dentist reads it

2. Data Aggregation and Insights: Centralized access to diagnostic patterns across your entire patient population enables identification of underdiagnosed conditions, regional health trends, and provider performance benchmarking impossible at the single-practice level

3. Operational Efficiency at Scale: The ROI math compounds across locations—reduced chair time per diagnosis, higher case acceptance through visual evidence, and fewer missed findings that become costly emergencies all multiply across your footprint

Expected Timeline: Decision to Full Deployment

Phase	Duration	Milestone
Pre-Implementation & Contracting	Weeks 1–2	Enterprise agreement signed, baseline metrics captured
Pilot Wave (2–3 locations)	Weeks 3–6	Pilot validation complete
Wave 2 Expansion (5–8 locations)	Weeks 7–10	Scaled playbook proven
Wave 3+ Full Deployment	Weeks 11–16	All locations live
Optimization & Stabilization	Weeks 17–20	ROI assessment complete

Total: 16–20 weeks from decision to full deployment for a 25-location DSO

---

2. Pre-Implementation Checklist (Weeks 1–2)

Technical Requirements

Hardware

☐ Verify all locations have workstations meeting minimum specs:

• Processor: Intel i5 (8th gen) or AMD Ryzen 5 equivalent or newer
• RAM: 16GB minimum (32GB recommended)
• Display: 1920x1080 resolution minimum; medical-grade monitors preferred for diagnostic work
• 🔵 Request official hardware requirements matrix from Dentem.ai

☐ Audit imaging equipment compatibility:

• Digital sensor types in use across locations
• Panoramic and CBCT units (make/model inventory)
• ⚠️ Legacy analog-to-digital converters may require replacement

Network

☐ Minimum bandwidth: 50 Mbps download / 25 Mbps upload per location ☐ Latency to Dentem.ai cloud infrastructure: <100ms preferred ☐ Verify firewall rules allow outbound HTTPS to Dentem.ai endpoints ☐ 🔵 Obtain IP whitelist requirements from vendor

Software

☐ Practice Management System version compatibility:

• Dentrix G7.1 or later
• Eaglesoft 21.0 or later
• Open Dental 22.1 or later ☐ Imaging software compatibility (Dexis, Apteryx, XDR, etc.) ☐ Browser requirements for web-based interface (if applicable) ☐ Operating system minimums (Windows 10 21H2 or later; Windows 11 supported)

Enterprise-Level Technical Requirements

Network Standards Across Locations

☐ 🟣 Decide: Centralized cloud hosting vs. hybrid edge deployment

• Cloud: Simplifies management, requires consistent WAN connectivity
• Hybrid: Faster local processing, more complex to maintain ☐ Document VPN/SD-WAN topology if applicable ☐ Verify consistent DNS and certificate management across locations

Single Sign-On (SSO)

☐ 🔵 Confirm Dentem.ai supports your identity provider (Okta, Azure AD, Google Workspace) ☐ Provision SSO integration environment ☐ Define role-based access levels:

• Corporate Admin (full configuration access)
• Regional Manager (multi-location read access, limited config)
• Location Admin (single-location configuration)
• Provider (clinical use only)
• Staff (view-only where applicable)

Centralized Credentialing

☐ Map Dentem.ai user accounts to provider NPI numbers ☐ Integrate with existing credentialing database if available ☐ Define automated provisioning/deprovisioning workflow for staff turnover

Vendor Onboarding Steps

Step	Owner	Timeline	Notes
☐ Execute enterprise BAA	Legal + Vendor	Day 1–3	🔵 Vendor-initiated
☐ Sign master services agreement	Legal + Procurement	Day 1–5
☐ Assign enterprise account manager	Vendor	Day 1	🔵
☐ Schedule technical kickoff call	IT + Vendor	Day 3–5	🔵
☐ Obtain sandbox/demo environment access	IT	Day 5–7	🔵
☐ Receive API documentation and credentials	IT	Day 5–7	🔵
☐ Confirm support SLAs and escalation contacts	Operations	Day 7	🔵

Key Vendor Contacts to Establish

☐ Enterprise Account Manager (primary relationship owner) ☐ Technical Implementation Lead (integration and configuration) ☐ Support Tier 2/3 escalation contact (for go-live issues) ☐ Customer Success Manager (post-implementation optimization)

Data/Access Prerequisites

☐ Generate API keys for PMS integration (per location or enterprise-level) ☐ Document imaging archive storage locations and access protocols ☐ Prepare sample imaging data from 2–3 locations for initial testing ☐ ⚠️ Verify patient consent language covers AI-assisted diagnosis (update if needed) ☐ Create service account credentials for automated workflows

Stakeholder Alignment Map

Stakeholder Level	Who	Communication Needed	Approval Required
Board/Investors	🟣	AI investment rationale, expected ROI, risk mitigation	Budget approval, strategic alignment
C-Suite (CEO, CFO, CDO)	🟣	Business case, timeline, resource requirements	Go/no-go decision, vendor selection
VP of Operations	Primary owner	Full playbook, weekly status, escalation authority	Wave advancement, rollback decisions
Chief Dental Officer		Clinical workflow changes, provider training plan, quality oversight	Clinical protocol approval
Regional Managers		Location readiness, rollout sequencing, champion identification	Local resource allocation
IT Director/Manager		Technical requirements, integration plan, security compliance	Architecture approval
Office Managers		Operational impact, training schedule, go-live logistics	Local scheduling
Providers (Dentists)		Clinical benefits, training requirements, workflow changes	None (inform and train)

Baseline Metrics to Capture

Clinical Metrics

Metric	Measurement Method	Capture Period
☐ Case acceptance rate	PMS treatment plan acceptance reports	90 days pre-implementation
☐ Average findings per FMX	Manual chart audit (sample 50 patients per location)	Point-in-time
☐ Time from imaging to treatment plan presentation	Timestamp analysis in PMS	30 days pre-implementation
☐ Re-treatment rate (e.g., missed caries requiring RCT)	Chart audit	12 months trailing

Operational Metrics

Metric	Measurement Method	Capture Period
☐ Average chair time per diagnostic appointment	PMS scheduling data	30 days pre-implementation
☐ Radiograph retake rate	Imaging system logs	30 days pre-implementation
☐ Provider time spent reviewing X-rays	Time study (sample 10 appointments per location)	Point-in-time

Financial Metrics

Metric	Measurement Method	Capture Period
☐ Average production per patient visit	PMS financial reports	90 days pre-implementation
☐ Insurance claim denial rate for diagnostic codes	RCM system	90 days pre-implementation
☐ Same-day treatment conversion rate	PMS treatment tracking	90 days pre-implementation

Standardizing Baseline Measurement Across Locations

☐ Create unified data dictionary defining each metric precisely ☐ 🟣 Require all locations to use identical reporting parameters ☐ Establish central data repository (spreadsheet or BI tool) for baseline capture ☐ Assign regional managers accountability for data completeness ☐ ⚠️ Address any PMS configuration variations that affect metric consistency before rollout

---

3. Location Readiness Assessment

Scoring Framework

Score each location on the following factors using a 1–5 scale:

Factor 1: IT Infrastructure Maturity

Score	Criteria
5	All hardware <3 years old, gigabit internet, current PMS version, no known IT issues
4	Hardware <5 years, 100+ Mbps internet, PMS within 2 versions of current
3	Mixed hardware ages, 50+ Mbps internet, some legacy equipment
2	Aging hardware requiring updates, inconsistent connectivity, PMS 3+ versions behind
1	Significant hardware refresh needed, unreliable internet, major PMS upgrade required first

Factor 2: Staff Tenure and Adaptability

Score	Criteria
5	Stable team (turnover <15%), prior successful tech adoption, enthusiastic about innovation
4	Low turnover, neutral-to-positive attitude toward new technology
3	Average turnover, mixed tech adoption history
2	Higher turnover (>30%), some resistance to recent changes
1	Unstable staffing, failed recent technology implementations, active change resistance

Factor 3: Patient Volume

Score	Risk/Impact Assessment
5	High volume (top quartile): Maximum ROI potential, highest implementation complexity
4	Above-average volume: Strong ROI, manageable complexity
3	Average volume: Balanced risk/reward for pilot
2	Below-average volume: Lower implementation risk, lower ROI impact
1	Low volume: Minimal impact, may not justify prioritization

Note: For pilot selection, prioritize scores of 3–4 (enough volume for meaningful validation without excessive risk)

Factor 4: Existing Tech Stack Compatibility

Score	Criteria
5	Confirmed compatible PMS + imaging system, existing integrations working well
4	Compatible systems, minor configuration expected
3	Compatible with known workarounds or middleware needed
2	Partial compatibility, significant integration work anticipated
1	Incompatible systems requiring replacement or major upgrade first

Factor 5: Local Champion Availability

Score	Criteria
5	Tech-forward provider + engaged office manager, both willing to lead
4	Strong champion available (either provider or office manager)
3	Potential champion identified but needs development
2	No clear champion, but no active resistance
1	No champion, potential resistors in key roles

Composite Scoring and Weighting

Factor	Weight	Rationale
IT Infrastructure Maturity	25%	Hard blocker if inadequate
Staff Tenure/Adaptability	20%	Critical for adoption success
Patient Volume	15%	ROI driver but not determinative
Tech Stack Compatibility	25%	Hard blocker if incompatible
Local Champion Availability	15%	Accelerator for successful rollout

Composite Score Calculation:

Score = (IT × 0.25) + (Staff × 0.20) + (Volume × 0.15) + (TechStack × 0.25) + (Champion × 0.15)

Readiness Classification

Composite Score	Classification	Rollout Recommendation
4.0–5.0	High Readiness	Wave 1 pilot candidates
3.0–3.9	Moderate Readiness	Wave 2 or early Wave 3
2.0–2.9	Low Readiness	Address gaps before rollout; late Wave 3
<2.0	Not Ready	Remediation required; defer until next phase

Sample Readiness Matrix

Location	IT (25%)	Staff (20%)	Volume (15%)	Tech (25%)	Champion (15%)	Composite	Wave
Denver Midtown	5	4	4	5	5	4.65	1
Austin South	4	4	3	4	4	3.85	1
Phoenix Central	4	3	5	4	3	3.80	1
Dallas Northwest	3	3	4	3	4	3.25	2
...	...	...	...	...	...	...	...

---

4. Rollout Strategy

Wave Structure Recommendation

Wave 1: Pilot (2–3 Locations)

Duration: 4 weeks Purpose: Validate integration, refine training materials, establish baseline performance

Wave 2: Controlled Expansion (5–8 Locations)

Duration: 4 weeks Purpose: Scale processes, stress-test support capacity, identify edge cases

Wave 3+: Full Deployment (Remaining Locations)

Duration: 4–6 weeks Purpose: Complete rollout using proven playbook

Wave 1 Pilot Location Selection Criteria

Select locations that are:

• ☐ High readiness (composite score 4.0+)
• ☐ Manageable risk (not your highest-revenue flagship locations)
• ☐ Representative of your broader portfolio (mix of specialties, markets, patient demographics)
• ☐ Geographically accessible for potential on-site troubleshooting
• ☐ Champion-rich (identified provider and OM willing to provide detailed feedback)

🟣 Recommendation: Include at least one location with a challenging element (e.g., older PMS version, higher-than-average turnover) to stress-test the playbook before scale.

Timeline Per Wave with Learning Capture

Wave 1 Timeline (Weeks 3–6)

Week	Activities
Week 3	Configuration, integration testing, champion training
Week 4	Staff training, parallel run begins
Week 5	Go-live, daily monitoring
Week 6	Issue resolution, feedback collection, Wave 1 retrospective
Buffer	1 week between Wave 1 and Wave 2 for playbook refinement

Wave 2 Timeline (Weeks 8–11)

Week	Activities
Week 8	Configuration using refined templates, champion training
Week 9	Staff training, parallel run
Week 10	Go-live (staggered: 2–3 locations per day)
Week 11	Stabilization, retrospective
Buffer	1 week for final playbook refinement

Wave 3+ Timeline (Weeks 13–16+)

Week	Activities
Weeks 13–14	Configuration and training (parallel across locations)
Weeks 15–16	Staggered go-live (4–6 locations per week)
Ongoing	Optimization as locations stabilize

Go/No-Go Criteria for Wave Advancement

Wave 1 → Wave 2 Requirements

Criterion	Threshold	Measurement
☐ Technical stability	<3 critical bugs outstanding	Vendor issue tracker
☐ Integration reliability	>99% successful image processing	System logs
☐ Staff adoption	>80% of trained staff using tool consistently	Usage analytics
☐ Provider satisfaction	Average rating ≥3.5/5	Post-pilot survey
☐ No HIPAA incidents	Zero reportable events	Compliance log
☐ Training materials validated	Champions confirm materials effective	Champion feedback

🟣 Decision Authority: VP of Operations (with CDO sign-off on clinical criteria)

Wave 2 → Wave 3 Requirements

All Wave 1 criteria, plus:

Criterion	Threshold	Measurement
☐ Support capacity validated	Average issue resolution <4 hours	Ticket tracking
☐ Scalable processes confirmed	No per-location customization required	Implementation team
☐ Early ROI indicators positive	Case acceptance trending ≥5% improvement	PMS reports

Rollback Plan

Triggers for Rollback Consideration

• Critical integration failure affecting patient care

• 10% of images failing to process after troubleshooting

• HIPAA/security incident
• Unified provider rejection (>50% refusing to use the tool)

Rollback Procedure

Step	Action	Timeline	Owner
1	🟣 Decision to pause made by VP Operations	Immediate	VP Ops
2	Notify vendor of pause and escalate support	Within 1 hour	IT Director
3	Disable Dentem.ai integration at affected locations	Within 2 hours	IT
4	Communicate to location staff: revert to pre-AI workflow	Within 2 hours	Regional Manager
5	Document all issues and root causes	Within 24 hours	Project Manager
6	Conduct root cause analysis with vendor	Within 48 hours	IT + Vendor 🔵
7	Determine remediation path and timeline	Within 1 week	All stakeholders

Critical: Rollback at one location does NOT automatically pause other waves unless issue is systemic.

---

5. Configuration & Integration (Weeks 2–3)

Practice Management System Integration

Dentrix Integration (Step-by-Step)

Step	Action	Time Est.	Owner
☐ 1	Verify Dentrix version compatibility (G7.1+)	15 min	Local IT
☐ 2	🔵 Obtain Dentrix API credentials from Dentem.ai	1–2 days	Vendor
☐ 3	Install Dentem.ai connector service on Dentrix server	30 min	IT
☐ 4	Configure connector with API credentials	15 min	IT
☐ 5	Map patient ID fields between systems	30 min	IT
☐ 6	⚠️ Configure imaging device routing (common failure point)	1 hour	IT + Vendor
☐ 7	Test with sample patient (non-production)	30 min	IT
☐ 8	Verify AI findings populate in clinical notes	15 min	IT + Provider
☐ 9	🔵 Complete integration validation checklist with vendor	1 hour	IT + Vendor

Eaglesoft Integration (Step-by-Step)

Step	Action	Time Est.	Owner
☐ 1	Confirm Eaglesoft 21.0+ and Patterson support agreement active	15 min	Local IT
☐ 2	🔵 Request Eaglesoft bridge configuration from Dentem.ai	1–2 days	Vendor
☐ 3	Enable third-party integration in Eaglesoft settings	15 min	IT
☐ 4	Install Dentem.ai bridge application	30 min	IT
☐ 5	Configure DICOM routing from imaging to Dentem.ai	45 min	IT
☐ 6	Map chart fields and clinical note templates	30 min	IT
☐ 7	⚠️ Test with multiple image types (BWX, PA, Pano)	1 hour	IT
☐ 8	Validate clinical note auto-population	15 min	Provider

Open Dental Integration (Step-by-Step)

Step	Action	Time Est.	Owner
☐ 1	Verify Open Dental 22.1+	10 min	Local IT
☐ 2	Enable API in Open Dental (Setup > Misc > API)	10 min	IT
☐ 3	🔵 Generate API key and share with Dentem.ai	15 min	IT + Vendor
☐ 4	Configure Dentem.ai with Open Dental endpoint	30 min	Vendor 🔵
☐ 5	Install image bridge (if not using native Open Dental imaging)	30 min	IT
☐ 6	Map procedure codes for diagnostic findings	30 min	IT
☐ 7	Test end-to-end workflow with sample patient	1 hour	IT + Provider

Imaging System Integration

Digital Sensor/Imaging Software Integration

Step	Action	Time Est.	Owner
☐ 1	Document imaging software type per location (Dexis, XDR, Apteryx, etc.)	2 hours	IT
☐ 2	🔵 Verify compatibility with Dentem.ai integration matrix	30 min	Vendor
☐ 3	Configure TWAIN or DICOM export settings	30 min/location	Local IT
☐ 4	⚠️ Set up automatic image routing to Dentem.ai	45 min/location	IT
☐ 5	Test image capture → AI analysis → result return workflow	30 min	IT + Provider
☐ 6	Verify image quality thresholds (reject low-quality before AI processing)	15 min	IT

CBCT Integration (if applicable)

Step	Action	Time Est.	Owner
☐ 1	Identify CBCT units (make/model) across locations	1 hour	IT
☐ 2	🔵 Confirm Dentem.ai CBCT module is included in contract	15 min	Vendor
☐ 3	Configure DICOM export from CBCT software	1 hour	IT
☐ 4	Test with sample scan (anonymized)	30 min	IT
☐ 5	Validate 3D rendering and annotation display	30 min	Provider

Test Environment Setup

Enterprise Test Environment Recommendation

🟣 Recommended Approach: Centralized test environment with location-specific test instances

Component	Approach	Rationale
Test PMS instance	Per-location (sandbox copies)	Mirrors production variability
Test imaging data	Centralized anonymized library	Consistent testing across locations
Dentem.ai test tenant	Single enterprise test tenant	Simplifies vendor support

Validation Checklist

Test Case	Expected Result	Pass/Fail
☐ Capture BWX and send to AI	AI analysis returns in <30 seconds
☐ Capture PA and send to AI	AI analysis returns in <30 seconds
☐ Capture Pano and send to AI	AI analysis returns in <90 seconds
☐ AI findings appear in PMS chart	Findings auto-populate clinical notes
☐ Provider modifies AI finding	Override saves correctly
☐ High-volume simulation (20 images in 10 min)	No queue backlog, all processed
☐ Network interruption recovery	Images queue locally, sync when restored
☐ User permission enforcement	Non-provider cannot modify findings

Data Migration / Historical Data Ingestion

Step	Action	Time Est.	Owner
☐ 1	🟣 Decide: Ingest historical images or start fresh	—	VP Ops + CDO
☐ 2	If ingesting: Identify date range (recommend 12–24 months)	30 min	Project Manager
☐ 3	Export historical images to staging folder	2–4 hours/location	IT
☐ 4	🔵 Upload to Dentem.ai batch processing queue	1–2 days	Vendor
☐ 5	⚠️ Validate patient ID matching (common source of errors)	2 hours	IT
☐ 6	Review AI findings on historical images (sample audit)	2 hours	Provider

Note: Historical ingestion is optional but valuable for establishing AI performance baselines and enabling retrospective quality review.

Enterprise Standardized Configuration Template

Settings to Standardize Centrally

Setting	Standard Value	Rationale
AI sensitivity threshold	Medium (vendor default)	Balance between catch-rate and false positives
Finding categories enabled	All (caries, bone loss, periapical, calculus, etc.)	Consistent diagnostic support
Annotation display	Always visible with toggle to hide	Provider preference flexibility
Report format	Standardized template (branded)	Consistent patient communication
Audit trail retention	7 years	HIPAA compliance
Alert thresholds	Critical findings → immediate pop-up	Patient safety

Settings Allowing Location-Specific Variation

Setting	Local Discretion	Rationale
Provider-specific display preferences	Color schemes, annotation styles	Individual workflow
Finding priority order	Customizable	Specialty mix (e.g., perio practice prioritizes bone loss)
Auto-populate vs. review-first workflow	Location choice	Provider comfort level
Patient report language	Adjustable	Demographics, health literacy

Enterprise HIPAA Compliance Checklist

Requirement	Action	Status
☐ Business Associate Agreement	🔵 Execute with Dentem.ai
☐ Data encryption in transit	Verify TLS 1.3 minimum
☐ Data encryption at rest	Verify AES-256
☐ Access logging	Confirm audit trail enabled
☐ User access controls	Implement role-based access per section 2
☐ Data retention policy	Align with organizational policy (minimum 6 years)
☐ Breach notification procedure	🔵 Document vendor's breach response SLA
☐ Employee training	Include AI tool in HIPAA training curriculum
☐ Risk assessment update	Add Dentem.ai to organizational risk assessment
☐ Subcontractor verification	🔵 Obtain Dentem.ai's subcontractor list and BAAs

---

6. Team Training Plan

Train-the-Trainer Model Overview

Central Training Team
        ↓ (certifies)
Location Champions (1 per location)
        ↓ (trains)
Location Staff (all roles)

Champion Selection Criteria

Criterion	Ideal Candidate Profile
Role	Office Manager or Lead Provider (or both working as a team)
Tenure	2+ years at location
Tech Aptitude	Demonstrated comfort with existing tech stack
Influence	Respected by peers, informal leadership
Availability	Capacity for 8–10 hours training/support during rollout
Attitude	Positive toward innovation, patient-centered mindset

Champion Responsibilities

Phase	Responsibility
Pre-Launch	Complete certification training; customize materials for local context
Launch Week	Deliver staff training; provide first-tier support; escalate issues
Post-Launch	Monitor adoption; conduct refresher sessions; onboard new hires
Ongoing	Participate in monthly champion calls; share best practices

Centralized Training Materials

Material	Created Centrally	Champion Customizes
Training video library	✓
Role-specific slide decks	✓
Day 1 cheat sheets	✓	Add location-specific contacts
Workflow integration guides	✓	Adjust for local workflow variations
FAQ document	✓	Add local context as questions arise
Patient communication scripts	✓	Localize language if needed

Role-Specific Training Outlines

Providers (Dentists) — Training Time: 90 minutes

Module	Duration	Format	Content
1. AI Overview & Evidence Base	15 min	Video	Clinical validation studies, FDA clearance, sensitivity/specificity data
2. Workflow Integration	30 min	Live demo	Where AI appears in clinical workflow, image capture → annotation → treatment planning
3. Interpreting AI Outputs	30 min	Hands-on	Reading annotations, confidence levels, finding categories, color coding
4. When to Override AI	10 min	Discussion	Clinical judgment remains paramount; documenting disagreement
5. Q&A	5 min	Live

Common Resistance Points & Responses:

Resistance	Response
"This will replace my clinical judgment"	"AI is a second set of eyes, not a replacement. You make all final decisions and the system documents your overrides."
"I don't trust the accuracy"	Share clinical validation data. Emphasize sensitivity/specificity rates. Offer parallel run period to build confidence.
"This will slow me down"	After initial learning curve, most providers report time savings. Workflow is designed for minimal clicks.

Day 1 Cheat Sheet — Providers:

╔════════════════════════════════════════════════════════════════╗
║ DENTEM.AI QUICK REFERENCE — PROVIDERS                          ║
╠════════════════════════════════════════════════════════════════╣
║ 1. AI activates automatically when image captured              ║
║ 2. Findings appear as colored overlays:                        ║
║    • Red = Caries • Blue = Bone loss • Yellow = Periapical    ║
║ 3. Click any finding for details + confidence %                ║
║ 4. To override: Click finding → "Dismiss" → Select reason     ║
║ 5. To add finding AI missed: Click "Add Finding" → Annotate   ║
║ 6. Findings auto-populate clinical notes — review before save ║
║ 7.