Implementation PlaybookDSO · Group Practice
Denti.AI
Step-by-step implementation guide — pre-implementation checklist, onboarding, staff training, go-live runbook, and ROI tracking.
Denti.AI — Implementation Playbook (DSO)
Denti.AI Implementation Playbook
Diagnostic Imaging AI for Dental Support Organizations
1. Executive Summary
What Denti.AI Does
Denti.AI is an FDA-cleared diagnostic imaging AI platform that automatically analyzes dental radiographs (bitewings, periapicals, panoramic, and CBCT images) to detect pathologies including caries, periapical lesions, bone loss, calculus, and other clinical findings. The platform integrates with existing imaging systems and practice management software to provide real-time AI-assisted annotations directly within the clinical workflow, enhancing diagnostic accuracy and documentation consistency.
Why DSOs Specifically Benefit from Diagnostic Imaging AI
Scale Advantages:
- Standardized diagnostic protocols across 15–50+ locations eliminate provider variability and reduce missed pathology rates at the enterprise level
- Centralized quality assurance becomes possible when every image is analyzed against the same AI standard
- Training and onboarding new providers accelerates when AI serves as a consistent diagnostic backstop
Standardization Benefits:
- Consistent documentation quality supports defensible treatment planning and reduces malpractice exposure across the organization
- Uniform case presentation increases case acceptance rates through objective, visual evidence shown to patients
- Standardized coding suggestions reduce claim denials and improve revenue cycle performance
Data Aggregation Opportunities:
- Enterprise-wide pathology prevalence data enables population health insights and strategic planning
- Cross-location benchmarking identifies high-performing and underperforming sites for targeted intervention
- Aggregated diagnostic data strengthens payer negotiations and supports value-based care initiatives
Expected Timeline
| Phase | Duration | Milestone |
|---|---|---|
| Decision to Contract Signed | 2–4 weeks | Legal, procurement, BAA executed |
| Pre-Implementation | 2 weeks | Technical readiness, stakeholder alignment |
| Pilot Wave (2–3 locations) | 4 weeks | Validated deployment, workflow refinement |
| Wave 2 Expansion (5–8 locations) | 4–6 weeks | Scaled deployment, champion network established |
| Wave 3+ Full Deployment | 6–10 weeks | Organization-wide go-live |
| Total: Decision to Full Deployment | 18–24 weeks | All locations live and optimized |
2. Pre-Implementation Checklist (Weeks 1–2)
Technical Requirements
Hardware Requirements per Location
☐ Workstations with minimum specs: Intel i5 (8th gen+) or AMD equivalent, 8GB RAM, SSD storage
☐ Display resolution minimum 1920x1080 (1440p+ recommended for imaging stations)
☐ Digital radiography sensors compatible with Denti.AI (verify sensor manufacturer list with vendor) 🔵
☐ Existing CBCT units verified for DICOM export compatibility (if applicable)
Software Requirements
☐ Operating system: Windows 10/11 Pro (64-bit) — verify macOS support timeline with vendor 🔵
☐ Browser requirements: Chrome 90+, Edge 90+ (for web-based interface)
☐ Practice management system version verification (see integration section)
☐ Imaging software version compatibility confirmed
Network Requirements
☐ Minimum bandwidth: 25 Mbps upload/download per location (50 Mbps+ recommended for high-volume sites)
☐ Latency under 100ms to Denti.AI cloud servers
☐ Firewall whitelist requirements obtained from vendor 🔵
☐ VPN compatibility confirmed (if applicable to your network architecture)
Enterprise-Level Network Standards 🟣
☐ Standardized network architecture decision: centralized image routing vs. location-direct upload
☐ Hosting model confirmed: Denti.AI cloud (standard) vs. on-premise option (if available) 🔵
☐ SSO integration requirements documented (SAML 2.0, Azure AD, Okta compatibility)
☐ Centralized credentialing system integration pathway defined
☐ Data residency requirements confirmed (HIPAA regions, state-specific requirements)
Vendor Onboarding Steps
| Step | Owner | Timeline | Notes |
|---|---|---|---|
| ☐ Execute BAA (Business Associate Agreement) 🔵 | Legal/Compliance | Week 1 | Non-negotiable HIPAA requirement |
| ☐ Master Services Agreement signed 🔵 | Procurement | Week 1 | Include SLA terms, support tiers |
| ☐ Obtain enterprise pricing confirmation 🔵 | Finance/Procurement | Week 1 | Per-location vs. per-image pricing model |
| ☐ Assign Denti.AI Customer Success Manager 🔵 | Vendor | Week 1 | Request dedicated DSO-experienced CSM |
| ☐ Establish technical implementation contact 🔵 | Vendor | Week 1 | Direct line for IT escalations |
| ☐ Schedule kick-off call with all stakeholders 🔵 | Project Lead | Week 1 | Include CDO, VP Ops, IT Director |
| ☐ Obtain implementation documentation package 🔵 | Vendor | Week 1 | API docs, integration guides, training materials |
Key Vendor Contacts to Establish
- Customer Success Manager (strategic relationship owner)
- Technical Implementation Specialist (integration and configuration)
- Support Escalation Contact (Tier 2+ issues)
- Executive Sponsor at Denti.AI (for executive-to-executive escalations)
Data/Access Prerequisites
☐ Create centralized Denti.AI admin account with enterprise permissions 🔵
☐ Document user provisioning workflow (manual vs. SSO-automated)
☐ ⚠️ Obtain API credentials for each PMS integration point (often takes longer than expected)
☐ Compile imaging archive access credentials for historical data ingestion (if desired)
☐ Create service accounts for system-to-system integrations (avoid tying to individual users)
☐ Document image storage architecture (local servers, cloud storage, or hybrid)
☐ Verify DICOM routing configuration capabilities at each location
Internal Stakeholder Alignment Map 🟣
Tier 1: Approval Required Before Proceeding
| Stakeholder | Role in Implementation | Key Concerns to Address | Communication Cadence |
|---|---|---|---|
| Board/Investors | Budget approval, strategic alignment | ROI timeline, competitive positioning, risk exposure | Quarterly board updates |
| CEO | Executive sponsorship | Strategic fit, resource allocation, timeline | Bi-weekly status briefings |
| Chief Dental Officer | Clinical validation, provider adoption | Clinical accuracy, workflow impact, liability | Weekly during implementation |
| CFO | Budget and ROI validation | Cost structure, payback period, accounting treatment | Monthly financial reviews |
Tier 2: Informed and Engaged
| Stakeholder | Role in Implementation | Key Concerns to Address | Communication Cadence |
|---|---|---|---|
| VP of Operations | Operational execution, rollout coordination | Workflow disruption, training burden, timeline | Weekly project status |
| IT Director | Technical implementation, security | Integration complexity, support burden, security | Daily during active phases |
| Regional Managers | Location coordination, change management | Staff impact, productivity dip, local resistance | Weekly regional syncs |
| Compliance Officer | HIPAA, regulatory oversight | BAA terms, data handling, audit trail | As-needed, pre-launch review |
Tier 3: Operational Execution
| Stakeholder | Role in Implementation | Key Concerns to Address | Communication Cadence |
|---|---|---|---|
| Office Managers | Local deployment leadership | Training time, schedule disruption, staff concerns | Daily during go-live week |
| Lead Providers | Clinical champions, peer influence | Workflow changes, AI accuracy, override process | Weekly during training |
| IT Support Staff | Technical troubleshooting | Support volume, escalation paths, documentation | Daily during active phases |
Baseline Metrics to Capture BEFORE Go-Live ⚠️
Critical: Standardize measurement methodology across all locations before capturing baselines
Diagnostic Efficiency Metrics
| Metric | Measurement Method | Source System | Capture Frequency |
|---|---|---|---|
| Average image-to-diagnosis time | Timestamp delta: image capture to treatment plan entry | PMS + Imaging software | Sample 50 cases/location |
| Radiographic findings documentation rate | % of images with documented pathology notes | PMS clinical notes | Full month audit |
| Re-take rate | % of images requiring re-capture | Imaging software logs | Monthly average |
Clinical Quality Metrics
| Metric | Measurement Method | Source System | Capture Frequency |
|---|---|---|---|
| Provider diagnostic concordance | Peer review of blinded cases across providers | Manual chart audit | Sample 20 cases/location |
| Missed pathology rate (if known) | Historical malpractice claims or peer review findings | Risk management records | Historical 12-month data |
| Patient return rate for missed findings | Patients returning within 6 months with pathology in previously imaged areas | PMS treatment history | 6-month retrospective |
Financial Metrics
| Metric | Measurement Method | Source System | Capture Frequency |
|---|---|---|---|
| Case acceptance rate | Accepted treatment value / Presented treatment value | PMS treatment planning | Monthly by location |
| Average treatment value per patient | Total production / Total unique patients | PMS production reports | Monthly average |
| Claim denial rate (diagnostic codes) | Denied claims / Total claims for diagnostic procedures | RCM system | Monthly average |
| Production per provider hour | Production / Clinical hours worked | PMS + Scheduling | Monthly average |
Standardization Protocol 🟣
☐ Create centralized data dictionary defining each metric precisely
☐ Deploy standardized reporting templates to all locations
☐ Designate a single analyst/team to compile baseline data
☐ Validate data quality from each location before accepting baselines
☐ Document any locations with incomplete or unreliable historical data
☐ Capture baselines 30 days before first pilot go-live
3. Location Readiness Assessment
Scoring Framework
Score each factor 1–5 per location. Calculate composite readiness score as weighted average.
Factor 1: IT Infrastructure Maturity (Weight: 25%)
| Score | Network Speed | Hardware Age | PMS Version | Overall Infrastructure |
|---|---|---|---|---|
| 5 | 100+ Mbps fiber | <2 years, enterprise-grade | Latest version, actively maintained | Cloud-first, standardized |
| 4 | 50–100 Mbps | 2–3 years, meets specs | Within 1 version of latest | Mostly standardized |
| 3 | 25–50 Mbps | 3–4 years, minimum specs | 2 versions behind | Mixed environment |
| 2 | 10–25 Mbps | 4–5 years, below specs | 3+ versions behind | Legacy systems present |
| 1 | <10 Mbps or unreliable | 5+ years, failing | Unsupported version | Significant technical debt |
Factor 2: Staff Tenure and Adaptability (Weight: 20%)
| Score | Turnover Rate | Tech Comfort | Training History | Change Readiness |
|---|---|---|---|---|
| 5 | <10% annual | Actively embraces new tech | Successfully adopted 2+ new systems in past year | Positive change culture |
| 4 | 10–20% annual | Comfortable with guided tech change | 1 successful adoption in past year | Generally receptive |
| 3 | 20–30% annual | Neutral, requires training | Mixed results with past changes | Average adaptability |
| 2 | 30–40% annual | Resistant but manageable | Struggled with recent changes | Below average |
| 1 | 40%+ annual | Actively resistant | Failed or abandoned recent implementations | High resistance risk |
Factor 3: Patient Volume (Weight: 15%)
| Score | Daily Patient Count | Strategic Consideration |
|---|---|---|
| 5 | 40+ patients/day | Highest ROI potential, highest risk — recommend for Wave 2 |
| 4 | 30–40 patients/day | Strong ROI, manageable risk — good Wave 1 candidate |
| 3 | 20–30 patients/day | Moderate volume, moderate impact |
| 2 | 10–20 patients/day | Lower volume, lower risk — good for learning |
| 1 | <10 patients/day | May not justify implementation effort |
Factor 4: Existing Tech Stack Compatibility (Weight: 25%)
| Score | PMS Compatibility | Imaging System | Integration History |
|---|---|---|---|
| 5 | Native Denti.AI integration available | Digital sensors fully compatible, DICOM-compliant | Clean API integrations, no workarounds |
| 4 | Verified compatible, standard integration | Digital sensors compatible, minor config needed | Mostly clean integrations |
| 3 | Compatible with custom configuration | Mixed digital/film, transitioning | Some integration workarounds |
| 2 | Limited compatibility, significant work needed | Older digital system, compatibility uncertain | Multiple workarounds required |
| 1 | Incompatible or requires PMS upgrade | Film-based or incompatible digital | Integration failures in past |
Factor 5: Local Champion Availability (Weight: 15%)
| Score | Champion Profile |
|---|---|
| 5 | Tech-forward provider AND engaged office manager, both volunteered for pilot |
| 4 | Strong provider or office manager champion, peer influence in region |
| 3 | Willing participant identified, moderate influence |
| 2 | Reluctant champion, assigned rather than volunteered |
| 1 | No champion identified, or identified champion is resistant |
Composite Score Calculation
Readiness Score = (Infrastructure × 0.25) + (Staff × 0.20) + (Volume × 0.15) + (Tech Stack × 0.25) + (Champion × 0.15)
Score Interpretation and Wave Assignment
| Composite Score | Readiness Tier | Recommended Wave | Notes |
|---|---|---|---|
| 4.5–5.0 | Tier 1: Pilot Ready | Wave 1 | Ideal pilot candidates |
| 3.5–4.4 | Tier 2: Ready | Wave 2 | Solid candidates after pilot learnings |
| 2.5–3.4 | Tier 3: Needs Preparation | Wave 3 | Address gaps before deployment |
| 1.5–2.4 | Tier 4: Significant Prep Required | Wave 4 or Hold | Major remediation required |
| 1.0–1.4 | Tier 5: Not Ready | Hold | Defer until fundamentals addressed |
Sample Location Assessment Matrix
| Location | Infrastructure | Staff | Volume | Tech Stack | Champion | Composite | Wave |
|---|---|---|---|---|---|---|---|
| Downtown Metro | 5 | 4 | 4 | 5 | 5 | 4.65 | Wave 1 |
| Suburban East | 4 | 5 | 3 | 4 | 4 | 4.10 | Wave 1 |
| Regional Center | 4 | 3 | 5 | 4 | 3 | 3.85 | Wave 2 |
| Community North | 3 | 4 | 3 | 3 | 4 | 3.35 | Wave 2 |
| Rural West | 2 | 3 | 2 | 2 | 2 | 2.20 | Hold |
Recommended Rollout Sequence Based on Scoring 🟣
Wave 1 Selection Criteria:
- Composite score ≥ 4.0
- No single factor score below 3
- At least one urban/suburban and one representative location type
- Geographic diversity for regional manager coverage
- Select 2–3 locations maximum
Wave 2 Selection Criteria:
- Composite score 3.5–4.4
- Address any single factor scores of 2 before go-live
- Include locations that stress-test scalability (higher volume)
- Select 5–8 locations
Wave 3+ Selection:
- All remaining Tier 3 locations after gap remediation
- Leverage champions from earlier waves for peer support
Deferred Locations:
- Create remediation plan with specific milestones
- Re-assess in 90 days
- Do not force deployment on unprepared locations
4. Rollout Strategy
Wave Structure Recommendation 🟣
Wave 1: Pilot Phase (2–3 Locations)
Duration: 4 weeks
Objective: Validate configuration, identify workflow issues, develop training refinements
Selection Criteria for Wave 1:
- Highest readiness scores (4.0+) but NOT highest volume (reduce risk exposure)
- Strong local champions who will provide detailed feedback
- Representative of broader portfolio (mix of practice types if applicable)
- Geographically convenient for on-site support visits
- Regional manager for these locations has bandwidth for intensive involvement
Wave 1 Timeline:
| Week | Activities |
|---|---|
| Week 1 | Configuration, integration testing, staff training |
| Week 2 | Soft go-live (limited patient load), daily monitoring |
| Week 3 | Full go-live, intensive support, issue documentation |
| Week 4 | Stabilization, workflow refinement, lessons learned capture |
Wave 2: Validated Scale (5–8 Locations)
Duration: 4–6 weeks
Objective: Prove scalability, stress-test support model, build champion network
Buffer Between Wave 1 and Wave 2: Minimum 2 weeks
- Consolidate Wave 1 learnings
- Update training materials based on real-world feedback
- Refine configuration templates
- Address any systemic issues discovered
Wave 2 Selection Criteria:
- Readiness score 3.5–4.4
- Include 1–2 higher-volume locations to test at scale
- Diverse geographic distribution to test support model
- Certified champions available from Wave 1 to assist
Wave 3+: Full Deployment (Remaining Locations)
Duration: 6–10 weeks (dependent on total location count)
Objective: Complete organizational deployment
Approach:
- Deploy 5–10 locations per week once Wave 2 validated
- Leverage Wave 1 and Wave 2 champions as peer mentors
- Standardized go-live process, minimal customization
- Central team in monitoring mode, escalation-only involvement
Go/No-Go Criteria Between Waves 🟣 ⚠️
Wave 1 → Wave 2 Go/No-Go (Executive Decision Required)
| Category | Go Criteria | No-Go Criteria |
|---|---|---|
| Technical Stability | <2 critical bugs unresolved, system uptime >99% | Any unresolved critical bugs affecting diagnosis |
| Workflow Integration | Staff completing workflows without workarounds | Fundamental workflow incompatibility identified |
| Provider Adoption | 80%+ providers actively using AI insights | Majority of providers bypassing or dismissing AI |
| Support Burden | Issues resolved within SLA, sustainable support load | Vendor SLAs missed, internal team overwhelmed |
| Clinical Accuracy | No false negatives on known pathology, false positive rate acceptable | Any missed critical pathology attributed to AI |
| Staff Satisfaction | Pulse survey average ≥ 3.5/5 | Pulse survey average < 3.0/5 |
Wave 2 → Wave 3 Go/No-Go
| Category | Go Criteria | No-Go Criteria |
|---|---|---|
| Scale Validation | Support model handled Wave 2 volume | Support model collapsed under Wave 2 load |
| Training Scalability | Champions successfully trained local teams | Champion model failed, central training required |
| ROI Indicators | Positive trends in case acceptance, documentation | No measurable improvement, negative trend |
| Location Variance | Performance consistent across Wave 2 sites | Significant unexplained variance between locations |
Rollback Plan ⚠️
Rollback Triggers
- Patient safety concern related to AI output
- System instability affecting practice operations
- Vendor unable to resolve critical issues within 72 hours
- Regulatory or compliance concern identified
Rollback Procedure
Location-Level Rollback:
- Champion notifies regional manager and central project lead
- Document specific failure and rollback justification
- Disable Denti.AI integration for affected location (retain data access)
- Revert to pre-Denti.AI imaging workflow
- Communicate to staff: "We're pausing AI assistance while we resolve [specific issue]"
- Location continues imaging normally without AI overlay
- Schedule root cause analysis within 48 hours
Wave-Level Pause:
- Regional managers escalate to VP Operations and Project Lead
- Executive decision to pause wave 🟣
- All in-progress deployments halted at current state
- Already-live locations continue (unless safety concern)
- Not-yet-live locations return to planning phase
- Vendor escalation to executive level 🔵
- Post-mortem and remediation plan before resuming
Important: Rollback is location-specific. A failure at one location does NOT automatically halt other waves unless systemic issue identified.
5. Configuration & Integration (Weeks 2–3)
Practice Management System Integration
Dentrix Enterprise Integration 🔵
| Step | Action | Owner | Est. Time |
|---|---|---|---|
| 1 | ☐ Verify Dentrix Enterprise version (20.0+ required) | IT | 30 min |
| 2 | ☐ Request Denti.AI Dentrix integration module 🔵 | Vendor | 1–2 days |
| 3 | ☐ Obtain API credentials from Henry Schein | IT | 3–5 days ⚠️ |
| 4 | ☐ Configure DICOM bridge for image routing | IT + Vendor | 2–4 hours |
| 5 | ☐ Install Denti.AI connector on Dentrix server | IT | 1 hour |
| 6 | ☐ Map patient identifiers between systems | IT + Vendor | 2–3 hours |
| 7 | ☐ Configure treatment code mapping for AI findings | Clinical + IT | 2–3 hours |
| 8 | ☐ Test bidirectional data flow with test patient | IT + Vendor | 1–2 hours |
| 9 | ☐ Validate findings appear in clinical notes | Clinical | 30 min |
Eaglesoft Integration 🔵
| Step | Action | Owner | Est. Time |
|---|---|---|---|
| 1 | ☐ Verify Eaglesoft version (21.0+ required) | IT | 30 min |
| 2 | ☐ Request Denti.AI Eaglesoft integration module 🔵 | Vendor | 1–2 days |
| 3 | ☐ Configure Patterson Bridge API access | IT | 2–3 hours |
| 4 | ☐ Install Denti.AI image capture listener | IT + Vendor | 1 hour |
| 5 | ☐ Configure automatic image routing to Denti.AI | IT | 1–2 hours |
| 6 | ☐ Map Eaglesoft patient IDs to Denti.AI | IT + Vendor | 2–3 hours |
| 7 | ☐ Test image upload and analysis return | IT + Vendor | 1–2 hours |
| 8 | ☐ Verify AI annotations visible in Eaglesoft imaging | Clinical | 30 min |
Open Dental Integration 🔵
| Step | Action | Owner | Est. Time |
|---|---|---|---|
| 1 | ☐ Verify Open Dental version (22.0+ recommended) | IT | 30 min |
| 2 | ☐ Enable Open Dental API (Program Links configuration) | IT | 1 hour |
| 3 | ☐ Generate API key for Denti.AI integration 🔵 | IT | 30 min |
| 4 | ☐ Configure Denti.AI with Open Dental API credentials | Vendor | 1 hour |
| 5 | ☐ Set up image bridge (varies by imaging software) | IT + Vendor | 2–4 hours |
| 6 | ☐ Test patient matching and image routing | IT + Vendor | 1–2 hours |
| 7 | ☐ Verify AI findings written back to patient chart | Clinical | 30 min |
Imaging System Integration
Digital Sensor/Phosphor Plate Systems 🔵
| Step | Action | Owner | Est. Time |
|---|---|---|---|
| 1 | ☐ Inventory all imaging hardware by location | IT | 2–4 hours total |
| 2 | ☐ Verify sensor compatibility with Denti.AI supported list 🔵 | IT + Vendor | 1 hour |
| 3 | ☐ Configure DICOM export settings on imaging software | IT | 1–2 hours/location |
| 4 | ☐ Set up image routing to Denti.AI (DICOM or proprietary) | IT + Vendor | 1–2 hours/location |
| 5 | ☐ Test image capture → analysis → return cycle | IT + Clinical | 30 min |
| 6 | ☐ Verify image quality meets AI minimum requirements | Clinical | 30 min |
Panoramic/Cephalometric Systems 🔵
| Step | Action | Owner | Est. Time |
|---|---|---|---|
| 1 | ☐ Verify panoramic unit DICOM compliance | IT | 30 min |
| 2 | ☐ Configure DICOM node for Denti.AI routing | IT | 1 hour |
| 3 | ☐ Test panoramic image analysis accuracy | Clinical + Vendor | 1 hour |
| 4 | ☐ Configure AI findings display for panoramic views | IT + Vendor | 30 min |
CBCT Systems (If Applicable) 🔵
| Step | Action | Owner | Est. Time |
|---|---|---|---|
| 1 | ☐ Verify CBCT unit compatibility (vendor-specific) 🔵 | IT + Vendor | 1 hour |
| 2 | ☐ Configure DICOM export for 3D datasets | IT | 2–3 hours |
| 3 | ☐ Verify CBCT analysis features included in contract 🔵 | Operations | 30 min |
| 4 | ☐ Test CBCT upload and analysis (significantly longer processing) | IT + Clinical | 2–3 hours |
| 5 | ☐ Train providers on 3D AI findings interpretation | Vendor + Clinical | 2 hours |
Test Environment Setup and Validation Checklist
Centralized Test Environment (Recommended for DSO) 🟣
☐ Deploy dedicated test instance of Denti.AI (non-production) 🔵
☐ Create test patient records with known pathology (gold standard images)
☐ Configure test PMS environment connected to test Denti.AI instance
☐ Establish test image dataset covering all pathology types Denti.AI detects
☐ Document expected AI findings for each test image (validation baseline)
Validation Checklist (Complete Before Each Location Go-Live)
| Test | Expected Result | Pass/Fail |
|---|---|---|
| ☐ Image upload latency | <5 seconds for intraoral, <30 seconds for pano | |
| ☐ AI analysis return time | <10 seconds for intraoral, <60 seconds for pano | |
| ☐ Caries detection accuracy | Detects all caries in test set | |
| ☐ Periapical lesion detection | Detects all periapical lesions in test set | |
| ☐ Bone loss measurement | Measurements within 1mm of reference | |
| ☐ PMS patient matching | Correct patient chart linked 100% | |
| ☐ Findings documentation | Findings appear in correct clinical notes section | |
| ☐ User access controls | Role-based access functions correctly | |
| ☐ Audit trail | All AI interactions logged appropriately |
Data Migration / Historical Ingestion
Historical Image Analysis (Optional) 🔵 🟣
| Step | Action | Owner | Est. Time |
|---|---|---|---|
| 1 | ☐ Determine scope: how far back to ingest (recommend 24 months) | Operations + Clinical | Decision |
| 2 | ☐ Verify historical image format compatibility 🔵 | IT + Vendor | 2–4 hours |
| 3 | ☐ Estimate image volume and ingestion timeline 🔵 | Vendor | 1–2 days |
| 4 | ☐ Negotiate historical ingestion pricing (may be separate) 🔵 | Finance | Variable |
| 5 | ☐ Schedule off-hours batch upload to avoid bandwidth impact | IT + Vendor | Plan |
| 6 | ☐ Run historical analysis in background (non-disruptive) | Vendor | Days to weeks |
| 7 | ☐ Review historical findings — clinical validation workflow | Clinical | Ongoing |
⚠️ Common Failure Point: Historical ingestion often reveals data quality issues (corrupted images, missing patient links, inconsistent naming). Budget extra time for data cleanup.
Security and HIPAA Compliance Verification
Enterprise HIPAA Checklist 🟣
| Requirement | Verification Step | Owner | Status |
|---|---|---|---|
| ☐ BAA executed | Confirm signed BAA on file, reviewed by legal | Compliance | |
| ☐ Data encryption in transit | Verify TLS 1.2+ for all data transmission 🔵 | IT + Vendor | |
| ☐ Data encryption at rest | Verify AES-256 or equivalent for stored images 🔵 | IT + Vendor | |
| ☐ Access controls | Role-based access configured, minimum necessary | IT | |
| ☐ Audit logging | All access to PHI logged, logs retained per policy | IT + Vendor | |
| ☐ User authentication | SSO or strong password policy enforced | IT | |
| ☐ Session management | Automatic timeout configured (recommend 15 min) | IT | |
| ☐ Incident response | Vendor incident response SLA documented 🔵 | Compliance + Vendor | |
| ☐ Data deletion procedures | Process for patient data deletion upon request 🔵 | Compliance + Vendor | |
| ☐ Subcontractor documentation | Denti.AI's subcontractors identified, BAAs confirmed 🔵 | Compliance | |
| ☐ Annual security review | Vendor provides SOC 2 Type II report 🔵 | Compliance |
Configuration Standardization
Standardize Centrally (Enterprise Configuration Template)
| Setting | Standard Configuration | Rationale |
|---|---|---|
| Detection thresholds | Vendor-recommended defaults | Consistency, clinical validation |
| Pathology categories enabled | All available categories | Comprehensive detection |
| Display overlay style | Vendor default with color coding | User familiarity |
| Documentation format | Structured findings with confidence scores | Defensibility, clarity |
| Alert sensitivity | Medium (vendor default) | Balance sensitivity/specificity |
| Integration polling frequency | Real-time | Minimize workflow disruption |
| User role definitions | Standard roles: Provider, Hygienist, Admin, View-Only | Consistent access control |
| Audit log retention | 7 years (match record retention policy) | Compliance alignment |
Allow Local Discretion (Location-Specific Configuration)
|
AI-generated implementation guide based on public vendor information. Verify specifics directly with Denti.AI.