How to Evaluate AI Clinical Tools for EMR Integration

What You'll Learn:

• 🎯 A comprehensive framework to evaluate AI clinical tools across security, interoperability, and workflow dimensions
• 📊 Specific technical requirements and integration standards for EMR compatibility
• ⚡ Critical security and compliance criteria that protect patient data and meet regulatory standards
• 💡 How to assess clinical workflow fit and measure real-world impact on physician burnout

The stakes have never been higher for clinical AI selection. With 63% of primary care physicians experiencing burnout and spending over 4 hours daily on EMR documentation, choosing the right AI clinical tool isn't just a technology decision—it's a clinical imperative that directly impacts physician wellbeing and patient care quality.

Yet most healthcare organizations lack a structured approach to evaluate AI clinical tools. Without a comprehensive assessment framework, you risk implementing solutions that create more problems than they solve: security vulnerabilities, integration nightmares, workflow disruptions, and ultimately, wasted investment and continued physician burnout.

📋 Executive Summary: The AI Clinical Tool Evaluation Framework

The challenge: Primary care physicians face an overwhelming array of AI clinical tools promising to reduce documentation burden and combat burnout. However, not all solutions deliver on their promises, and selecting the wrong tool can introduce new complexities into already strained clinical workflows.

The solution: A systematic evaluation framework that assesses AI clinical tools across five critical dimensions: security and compliance, EMR integration capabilities, clinical workflow fit, technical architecture, and measurable clinical outcomes.

Key capabilities to evaluate:

• Security posture: HIPAA compliance, SOC 2 Type II certification, encryption standards, and access controls
• Integration depth: Support for major EMR systems (Epic, Cerner, athenahealth), integration methods (HL7, FHIR, SMART on FHIR), and bi-directional data flow
• Workflow intelligence: Proactive vs. reactive capabilities, clinical decision support, and orchestration beyond documentation
• Implementation complexity: Typical deployment timeline (5-10 days for advanced systems), technical requirements, and support infrastructure

Critical differentiation: Traditional AI scribes reduce documentation time by 4% but fail to address the root cause of physician burnout—the cognitive burden of managing complex clinical workflows. Conversational Clinical Operating Systems represent the next evolution, orchestrating entire clinical workflows and delivering 13% burnout reduction in 30 days.

This guide provides a comprehensive framework to evaluate AI clinical tools systematically, ensuring your organization selects a solution that delivers measurable clinical value while meeting rigorous security and integration standards.

---

🏗️ Understanding AI Clinical Tool Architecture

Before you can effectively evaluate AI clinical tools, you need to understand the fundamental architectural differences that separate reactive documentation tools from proactive clinical operating systems.

System Architecture: Reactive vs. Proactive AI

The architecture of an AI clinical tool determines its capabilities and limitations. Most AI scribes operate on a simple linear architecture, while advanced conversational clinical operating systems employ a multi-layered, orchestration-based approach.

Component Breakdown: What Powers Clinical AI

Understanding the core components helps you evaluate AI clinical tools more effectively and ask the right technical questions during vendor assessments.

Component	Reactive AI Scribes	Proactive Clinical OS
Speech Recognition	Medical-grade ASR	Medical-grade ASR + contextual understanding
Natural Language Processing	Basic transcription	Advanced clinical entity extraction
Clinical Intelligence	Template-based documentation	Predictive analytics + clinical reasoning
EMR Integration	One-way documentation push	Bi-directional data synchronization
Workflow Engine	None	Multi-step orchestration across EMR functions
Decision Support	None	Real-time clinical guidelines integration

Data Flow and Integration Points

The data flow architecture reveals how deeply an AI clinical tool integrates with your existing systems and whether it can truly orchestrate clinical workflows or merely document them.

Reactive AI Scribe Data Flow:

1. Audio capture during patient encounter
2. Transcription and basic clinical note generation
3. One-way push to EMR documentation field
4. Physician manually completes remaining workflow tasks

Proactive Clinical OS Data Flow:

1. Audio capture with real-time contextual analysis
2. Bi-directional EMR data retrieval (patient history, active medications, recent labs)
3. Multi-output generation: documentation + orders + forms + clinical decision support
4. Intelligent orchestration of next actions based on clinical context
5. Continuous learning from physician preferences and feedback

Technology Stack Considerations

When you evaluate AI clinical tools, understanding the underlying technology stack helps assess scalability, security, and long-term viability.

Critical technology stack components:

• AI/ML frameworks: Modern large language models (LLMs) fine-tuned for clinical applications vs. older rule-based systems
• Cloud infrastructure: HIPAA-compliant cloud platforms (AWS, Azure, Google Cloud) with healthcare-specific configurations
• API architecture: RESTful APIs with robust authentication for seamless EMR integration
• Database systems: Encrypted, redundant databases designed for healthcare data residency requirements
• Security layers: Multi-factor authentication, role-based access control, end-to-end encryption

Key question to ask vendors: "What specific AI models power your clinical intelligence, and how do you ensure they stay current with evolving clinical guidelines?"

---

🔗 EMR Integration: The Foundation of Clinical AI Success

EMR integration capabilities determine whether an AI clinical tool becomes a seamless part of your workflow or another disconnected system that increases cognitive burden. When you evaluate AI clinical tools, integration depth matters far more than integration breadth.

Supported EMR Systems and Integration Depth

Not all EMR integrations are created equal. A vendor claiming "Epic integration" might only push documentation to a single field, while comprehensive integration orchestrates orders, forms, referrals, and clinical decision support across multiple EMR modules.

EMR System	Market Share	Integration Complexity	Typical Capabilities
Epic	31%	High	Deep API access, SMART on FHIR, custom workflows
Cerner (Oracle Health)	25%	High	HL7 integration, limited API access
athenahealth	6%	Medium	API-first architecture, good documentation
MEDITECH	14%	Medium	HL7 primary, emerging FHIR support
eClinicalWorks	4%	Medium	API available, variable implementation
AllScripts	3%	Medium	HL7 and API options

Critical evaluation criteria:

• Read access: Can the AI tool retrieve patient history, medications, allergies, and recent lab results to inform clinical decision support?
• Write access: Does it only push documentation, or can it also create orders, populate forms, and initiate referrals?
• Real-time synchronization: Are updates instantaneous or delayed by batch processing?
• Module coverage: How many EMR modules can it interact with (documentation, orders, labs, imaging, referrals, billing)?

Integration Methods: Technical Standards That Matter

Understanding integration methods helps you evaluate AI clinical tools for long-term sustainability and interoperability. Legacy systems relying solely on screen scraping or manual data entry represent significant technical debt.

HL7 v2 Integration:

• Pros: Widely supported, mature standard, reliable for basic data exchange
• Cons: Limited semantic interoperability, complex message parsing, primarily one-way communication
• Best for: Basic ADT (Admit/Discharge/Transfer) data, lab results, simple documentation

FHIR (Fast Healthcare Interoperability Resources):

• Pros: Modern RESTful API standard, rich semantic definitions, growing adoption
• Cons: Still emerging in some EMR systems, requires robust API infrastructure
• Best for: Bi-directional data exchange, complex clinical workflows, patient-generated data

SMART on FHIR:

• Pros: Combines FHIR data access with OAuth 2.0 security, EMR-agnostic authentication
• Cons: Requires EMR support, implementation varies by vendor
• Best for: Third-party applications that need secure, standardized EMR access

Direct API Integration:

• Pros: Deepest integration possible, custom workflows, real-time performance
• Cons: EMR-specific development, requires ongoing maintenance as APIs evolve
• Best for: Enterprise deployments requiring maximum functionality

Data Synchronization and Bi-Directional Updates

The difference between reactive and proactive AI becomes most apparent in data synchronization capabilities. When you evaluate AI clinical tools, assess not just whether they integrate with your EMR, but how intelligently they orchestrate data flow.

One-way synchronization (AI Scribes):

• AI tool pushes documentation to EMR
• No retrieval of patient context
• No awareness of existing orders or clinical state
• Physician manually completes remaining workflow

Bi-directional synchronization (Clinical Operating Systems):

• AI tool retrieves patient history, medications, allergies, recent labs
• Analyzes clinical context to suggest appropriate orders
• Updates EMR across multiple modules (documentation, orders, forms)
• Continuously syncs to reflect physician modifications

Key evaluation questions:

1. "Can your AI tool retrieve my patient's active medication list and recent lab results before suggesting orders?"
2. "If I modify an AI-generated order, does the system learn from that preference for future encounters?"
3. "How does your system handle conflicting data between what I say verbally and what exists in the EMR?"
4. "What happens if EMR connectivity is temporarily lost during a patient encounter?"

Custom Integration Support and Implementation

Healthcare organizations operate in diverse technical environments. When you evaluate AI clinical tools, assess the vendor's ability to support custom integration requirements and legacy systems.

Enterprise integration considerations:

• Legacy system support: Can the AI tool integrate with older EMR versions or custom-built systems?
• Multi-EMR environments: If your organization uses different EMRs across locations, can the AI tool support multiple systems simultaneously?
• Custom workflows: Can integration be tailored to your organization's specific clinical workflows and documentation requirements?
• Interface engine compatibility: Does the AI tool work with your existing interface engine (e.g., Rhapsody, Mirth, Corepoint)?

Implementation support to evaluate:

• Integration specialists: Does the vendor provide dedicated technical resources for complex integrations?
• Testing environments: Are sandbox environments available for thorough testing before production deployment?
• Documentation quality: Is technical integration documentation comprehensive and current?
• Ongoing maintenance: Who manages integration updates when EMR systems are upgraded?

For more context on how proactive clinical AI differs from reactive documentation tools, explore the fundamental architectural differences that enable true workflow orchestration.

---

🔒 Security and Compliance: Non-Negotiable Requirements

Security and compliance aren't just checkboxes when you evaluate AI clinical tools—they're foundational requirements that protect patient data, ensure regulatory adherence, and mitigate organizational risk. A security breach can cost healthcare organizations an average of $10.93 million per incident (IBM Security, 2025), making this evaluation dimension critical.

HIPAA Compliance: Beyond the Basics

Every AI clinical tool vendor claims HIPAA compliance, but the depth of that compliance varies dramatically. Comprehensive HIPAA compliance extends beyond basic technical safeguards to encompass administrative, physical, and technical controls.

HIPAA compliance evaluation checklist:

Requirement	Minimum Standard	Advanced Implementation
Business Associate Agreement	Standard BAA template	Comprehensive BAA with specific breach notification timelines
Access Controls	Basic username/password	Multi-factor authentication + role-based access + session timeouts
Audit Logging	Basic activity logs	Comprehensive audit trails with tamper-proof storage
Data Encryption	Encryption at rest	Encryption at rest + in transit + in use (processing)
Breach Notification	Within 60 days	Immediate notification + detailed forensics
Employee Training	Annual HIPAA training	Continuous security training + phishing simulations

Critical questions to ask vendors:

1. "Can you provide evidence of your HIPAA compliance program, including recent risk assessments and security reviews?"
2. "What is your process for handling potential security incidents, and what is your breach notification timeline?"
3. "How do you ensure subcontractors and third-party services are also HIPAA compliant?"
4. "Can you provide references from other healthcare organizations regarding your security practices?"

SOC 2 Type II Certification and Third-Party Audits

SOC 2 Type II certification demonstrates that an AI clinical tool vendor has implemented comprehensive security controls and undergone independent third-party auditing. This certification is increasingly becoming table stakes for healthcare AI vendors.

SOC 2 Trust Service Criteria:

• Security: Protection against unauthorized access (physical and logical)
• Availability: System uptime and operational performance
• Processing Integrity: System processing is complete, valid, accurate, timely, and authorized
• Confidentiality: Information designated as confidential is protected
• Privacy: Personal information is collected, used, retained, disclosed, and disposed of in conformity with commitments

Why SOC 2 Type II matters:

• Type I vs. Type II: Type I audits design of controls at a point in time; Type II audits operating effectiveness over 6-12 months
• Independent verification: Third-party auditors validate security claims
• Continuous improvement: Annual audits require ongoing security program maturity
• Customer assurance: Demonstrates commitment to security beyond marketing claims

Additional certifications to evaluate:

• HITRUST CSF Certification: Healthcare-specific security framework combining multiple standards
• ISO 27001: International standard for information security management
• FedRAMP: Required for AI tools serving federal healthcare facilities

Data Encryption Standards: At Rest, In Transit, and In Use

Encryption is the last line of defense protecting patient data. When you evaluate AI clinical tools, scrutinize encryption implementations across all data states.

Encryption requirements by data state:

At Rest Encryption:

• Standard: AES-256 encryption for all stored data
• Key management: Hardware Security Modules (HSMs) or cloud-native key management services
• Backup encryption: Encrypted backups with separate encryption keys
• Database encryption: Transparent Data Encryption (TDE) for database-level protection

In Transit Encryption:

• Standard: TLS 1.3 or higher for all network communications
• Certificate management: Valid, non-expired SSL/TLS certificates from trusted authorities
• API security: Encrypted API calls with token-based authentication
• EMR integration: Encrypted channels for all EMR data exchange

In Use Encryption (Emerging):

• Memory encryption: Protecting data while being processed in memory
• Confidential computing: Hardware-based trusted execution environments
• Homomorphic encryption: Processing encrypted data without decryption (future state)

Access Controls and Authentication

Robust access controls ensure that only authorized users can access patient data, and that access is limited to the minimum necessary for clinical functions.

Multi-layered access control framework:

1. Authentication: Verifying user identity

   • Multi-factor authentication (MFA) required for all users
   • Single Sign-On (SSO) integration with organizational identity providers
   • Biometric authentication options for mobile devices

2. Authorization: Determining what authenticated users can access

   • Role-Based Access Control (RBAC): Permissions based on clinical role
   • Attribute-Based Access Control (ABAC): Permissions based on user attributes, resource attributes, and environmental conditions
   • Principle of least privilege: Users receive minimum access necessary

3. Session Management:

   • Automatic session timeouts after inactivity (typically 15-30 minutes)
   • Concurrent session limits
   • Secure session token management

4. Access Monitoring:

   • Real-time monitoring of access patterns
   • Anomaly detection for unusual access behavior
   • Regular access reviews and recertification

Evaluation questions:

• "How does your system integrate with our existing identity management infrastructure?"
• "Can we configure role-based access controls to match our organizational structure?"
• "What monitoring and alerting capabilities exist for suspicious access patterns?"

Audit Logging and Compliance Reporting

Comprehensive audit logging enables security monitoring, incident investigation, and regulatory compliance. When you evaluate AI clinical tools, assess the depth and accessibility of audit capabilities.

Essential audit log elements:

Log Component	Required Information	Retention Period
User Actions	User ID, action type, timestamp, affected resources	7 years (HIPAA requirement)
Data Access	User ID, patient record accessed, fields viewed, timestamp	7 years
System Changes	Administrator actions, configuration changes, timestamp	7 years
Authentication	Login attempts (successful and failed), MFA events, logout	7 years
Integration Events	EMR data exchange, API calls, integration errors	7 years
Security Events	Access violations, suspicious activity, security alerts	7 years

Audit log security requirements:

• Tamper-proof storage: Logs stored in write-once, read-many (WORM) storage or blockchain-based immutable logs
• Separate access controls: Audit logs protected with different access controls than operational data
• Real-time monitoring: Security Information and Event Management (SIEM) integration for real-time alerting
• Searchability: Efficient search and filtering capabilities for compliance reporting and incident investigation

Compliance reporting capabilities:

• Pre-built reports for common compliance requirements (HIPAA, HITECH, state regulations)
• Custom report builder for organization-specific needs
• Automated compliance monitoring with alerts for potential violations
• Export capabilities for external auditors

Understanding how clinical workflow automation intersects with security requirements helps evaluate whether AI tools maintain security while improving efficiency.

---

🚀 Implementation Guide: From Evaluation to Production

A comprehensive implementation process distinguishes successful AI clinical tool deployments from failed projects. When you evaluate AI clinical tools, assess not just the technology but the vendor's implementation methodology and support infrastructure.

Pre-Implementation Requirements and Readiness Assessment

Successful implementation begins long before deployment. A structured readiness assessment identifies potential obstacles and ensures your organization is prepared for change.

Technical readiness checklist:

• Infrastructure assessment:

  • Network bandwidth sufficient for real-time audio streaming (minimum 1 Mbps per concurrent user)
  • Firewall rules configured to allow AI tool communication
  • EMR API access credentials and permissions secured
  • Test environment availability for integration validation

• Security readiness:

  • Business Associate Agreement (BAA) executed
  • Security review completed and approved
  • Data residency requirements documented
  • Incident response procedures updated to include AI tool

• Clinical readiness:

  • Champion physicians identified for pilot phase
  • Clinical workflows documented and analyzed
  • Success metrics defined and baseline measurements captured
  • Training schedule developed for physician and staff education

Organizational readiness assessment:

Installation and Configuration Process

Implementation complexity varies dramatically across AI clinical tools. Advanced Conversational Clinical Operating Systems typically deploy faster than legacy AI scribes due to modern API-based architectures.

Typical implementation timeline:

Phase	Duration	Key Activities	Success Criteria
Planning	3-5 days	Kickoff meeting, requirements gathering, technical discovery	Implementation plan approved
Configuration	2-3 days	System configuration, EMR integration setup, security configuration	Test environment functional
Testing	2-3 days	Integration testing, workflow validation, security testing	All test cases passed
Training	1-2 days	Physician training, staff training, workflow education	90%+ training completion
Pilot	5-10 days	Limited physician group, close monitoring, rapid iteration	Positive physician feedback
Go-Live	1 day	Full deployment, support availability, monitoring	Successful first encounters
Optimization	Ongoing	Workflow refinement, feedback integration, performance tuning	Sustained adoption metrics

Configuration steps for comprehensive evaluation:

1. EMR integration configuration:

   • API credentials and authentication setup
   • Data mapping between AI tool and EMR fields
   • Bi-directional synchronization rules
   • Error handling and retry logic

2. Clinical workflow customization:

   • Documentation templates aligned with organizational standards
   • Order set integration and favorites configuration
   • Clinical decision support rules customization
   • Specialty-specific workflow adjustments

3. User provisioning and access control:

   • User accounts created with appropriate roles
   • Single Sign-On (SSO) integration configured
   • Multi-factor authentication enabled
   • Access policies aligned with organizational security requirements

4. Device and hardware setup:

   • Microphone hardware selection and configuration
   • Mobile device deployment (if applicable)
   • Exam room equipment installation
   • Network connectivity verification

Testing Protocols and Quality Assurance

Rigorous testing prevents production issues and ensures the AI clinical tool performs as expected in real clinical workflows. When you evaluate AI clinical tools, assess the vendor's testing methodology and quality assurance processes.

Comprehensive testing framework:

Integration Testing:

• EMR data retrieval accuracy (patient demographics, medications, allergies, labs)
• Documentation push reliability and formatting
• Order entry functionality across all order types
• Bi-directional synchronization validation
• Error handling and recovery procedures

Clinical Accuracy Testing:

• Medical terminology recognition accuracy
• Clinical entity extraction precision (medications, dosages, diagnoses)
• Order suggestion relevance and appropriateness
• Clinical decision support accuracy
• Specialty-specific terminology handling

Performance Testing:

• Response time for documentation generation
• System performance under concurrent user load
• Audio quality and transcription accuracy in various environments
• Mobile device performance and battery consumption
• Network resilience and offline capabilities

Security Testing:

• Penetration testing results
• Vulnerability scanning reports
• Authentication and authorization validation
• Encryption verification
• Audit logging completeness

User Acceptance Testing (UAT):

• Physician workflow validation with real clinical scenarios
• Staff workflow integration testing
• Usability assessment and feedback collection
• Edge case handling (complex patients, unusual presentations)

Go-Live Checklist and Production Deployment

A structured go-live process minimizes disruption and ensures rapid issue resolution. When you evaluate AI clinical tools, review the vendor's go-live support model.

Go-live checklist:

• All integration testing completed successfully
• Security review approved and documented
• Training completion verified for all users
• Support escalation procedures communicated
• Monitoring and alerting configured
• Rollback plan documented and tested
• Success metrics baseline established
• Communication plan executed (physicians, staff, patients)
• Vendor support team on standby for go-live support
• Leadership informed and available for issue escalation

First-day support model:

• Onsite support: Vendor technical resources present for immediate issue resolution
• Dedicated support channel: Direct access to technical support via phone, chat, or video
• Rapid response commitment: <15 minute response time for critical issues
• Real-time monitoring: Vendor monitoring system performance and proactively addressing issues
• Daily debriefs: End-of-day review sessions to address feedback and plan optimizations

Implementation Timeline: Realistic Expectations

Implementation timelines vary based on organizational complexity, EMR system, and integration depth. When you evaluate AI clinical tools, set realistic expectations based on deployment complexity.

Timeline comparison by tool type:

Tool Type	Typical Timeline	Integration Depth	Configuration Complexity
Basic AI Scribe	2-4 weeks	Shallow (documentation only)	Low
Advanced AI Scribe	3-6 weeks	Medium (documentation + basic orders)	Medium
Clinical Operating System	5-10 days	Deep (full workflow orchestration)	High initial, streamlined process

Why advanced systems can deploy faster:

Modern conversational clinical operating systems leverage API-first architectures and pre-built EMR integrations, enabling faster deployment despite deeper functionality. Legacy AI scribes often require custom integration work for each deployment.

Factors that extend implementation timelines:

• Complex EMR customizations requiring custom integration development
• Multi-location deployments with varying workflows
• Extensive security reviews in highly regulated environments
• Legacy EMR versions with limited API capabilities
• Organizational change management challenges

For insights on reducing EMR documentation burden during implementation, review strategies for reducing EMR documentation time.

---

💻 Technical Requirements and Infrastructure Considerations

Understanding technical requirements helps you evaluate AI clinical tools for compatibility with your existing infrastructure and anticipate necessary investments. Inadequate infrastructure planning is a leading cause of implementation delays and performance issues.

Infrastructure Requirements Overview

Modern AI clinical tools operate as cloud-based services, minimizing on-premises infrastructure requirements. However, network, device, and security infrastructure considerations remain critical.

Core infrastructure requirements:

Component	Minimum Requirement	Recommended	Enterprise
Network Bandwidth	1 Mbps per user	5 Mbps per user	10+ Mbps per user
Network Latency	<200ms to cloud service	<100ms	<50ms
Internet Connection	Standard broadband	Redundant connections	Multiple redundant with automatic failover
Firewall Rules	Allow HTTPS (443) to AI service	Dedicated firewall rules with monitoring	Application-aware firewall with DPI
WiFi Coverage	Basic coverage in exam rooms	Enterprise-grade WiFi with roaming	WiFi 6 with seamless roaming and QoS
Backup Connectivity	None	4G/5G backup	Automatic failover to cellular backup

Network Requirements and Performance Considerations

Network performance directly impacts AI clinical tool responsiveness and physician satisfaction. When you evaluate AI clinical tools, assess network requirements against your current infrastructure.

Network performance requirements:

Critical network considerations:

• Bandwidth requirements:

  • Audio streaming: 64-128 kbps per concurrent user
  • API data exchange: 256-512 kbps per concurrent user
  • Total per user: 1-2 Mbps recommended for optimal performance

• Latency sensitivity:

  • <100ms latency: Excellent real-time responsiveness
  • 100-200ms latency: Acceptable for most clinical workflows

  • 200ms latency: Noticeable delays, potential physician frustration

• Quality of Service (QoS):

  • Prioritize AI clinical tool traffic over non-critical applications
  • Implement traffic shaping to ensure consistent performance
  • Monitor network performance and adjust QoS policies as needed

• Network security:

  • Firewall rules allowing encrypted traffic to AI service endpoints
  • Intrusion detection/prevention systems configured for AI tool traffic patterns
  • Network segmentation separating clinical systems from general network

Network assessment questions for vendors:

1. "What are your service's IP addresses or domain names for firewall configuration?"
2. "What is your service's uptime SLA and historical uptime performance?"
3. "How does your system handle network interruptions or degraded connectivity?"
4. "What network monitoring do you recommend to ensure optimal performance?"

Browser Compatibility and Device Support

AI clinical tools must support diverse device ecosystems in modern healthcare environments. When you evaluate AI clinical tools, verify compatibility with your organization's device standards.

Browser compatibility matrix:

Browser	Minimum Version	Recommended	Notes
Google Chrome	90+	Latest	Best performance, full feature support
Microsoft Edge	90+ (Chromium)	Latest	Chromium-based Edge fully supported
Safari	14+	Latest	macOS and iOS support
Firefox	88+	Latest	Full support with occasional audio quirks
Internet Explorer	Not supported	N/A	Legacy browser, security concerns

Device support categories:

Desktop/Laptop:

• Windows 10/11 with supported browsers
• macOS 11+ with supported browsers
• Linux (Ubuntu, CentOS) with Chrome/Firefox
• Minimum 4GB RAM, 8GB recommended

Mobile Devices:

• iOS 14+ (iPhone, iPad)
• Android 10+ (phone, tablet)
• Native