Skip to main content

AI Transparency Disclosure

We believe you have the right to know how artificial intelligence is used in the tools you trust with your patients' care. This page provides complete transparency about our AI features.

Last Updated: 01/07/2026

HIPAA CompliantNYC A-820 CompliantONC DSI Ready

Our Core Commitment

We do not use your patient data to train AI models. All AI processing is performed by HIPAA-compliant providers bound by Business Associate Agreements. Your data is used only to provide services to you—never for model training, never sold, never shared for advertising.

AI Features in Joyera

Each AI feature below includes details about its purpose, the data it uses, how models are trained, and human oversight mechanisms.

Clinical Documentation Assistant

Purpose

Generates draft clinical notes, SOAP documentation, IEP progress notes, and therapy session summaries from provider dictation or session transcripts.

Data Used

  • Audio recordings or transcripts of clinical sessions (with patient/guardian consent)
  • Patient demographics and clinical history
  • Provider preferences and documentation style
  • Billing codes and service types

Training Data

Uses commercial AI models (commercial AI models) trained on publicly available clinical documentation examples and medical literature. No Joyera customer data is used to train these models.

Human Oversight

All AI-generated documentation requires review and approval by a licensed healthcare professional before being finalized. The provider has full editing capability and bears sole responsibility for the final clinical record.

Demographic Factors

Age and diagnosis may influence documentation templates and suggested ICD-10 codes. No race, ethnicity, or socioeconomic factors influence clinical suggestions.

Billing Code Suggestion

Purpose

Suggests appropriate CPT, ICD-10, and HCPCS codes based on clinical documentation to support accurate medical billing.

Data Used

  • Clinical session notes and documentation
  • Service type and duration
  • Patient diagnosis and treatment goals
  • Payer requirements and coverage rules

Training Data

Uses commercial AI models trained on publicly available medical coding guidelines (CMS, AMA CPT manuals). Code suggestions are based on official coding rules, not historical billing patterns.

Human Oversight

All code suggestions are presented as recommendations only. Providers must verify and select appropriate codes. The billing professional or provider is solely responsible for code accuracy and compliance.

AI Chat Assistant (Joy)

Purpose

Provides conversational assistance for scheduling, patient lookup, task management, billing inquiries, and general practice workflow support.

Data Used

  • Conversation history within the current session
  • Practice calendar and appointment data
  • Patient demographics (name, DOB, contact info)
  • Task and billing status information

Training Data

Uses commercial AI models (commercial AI models) with Joyera-specific prompts for healthcare workflows. No customer data is used for model training. Practice-specific knowledge (protocols, SOPs) may be retrieved from practice knowledge base when enabled.

Human Oversight

Users can review all AI responses, request clarification, or override AI suggestions. All actions (scheduling, task creation) require explicit user confirmation before execution.

Transcription Services

Purpose

Converts audio recordings of clinical sessions into text transcripts for documentation purposes.

Data Used

  • Audio recordings of clinical sessions
  • Speaker identification metadata
  • Session date and time

Training Data

Uses commercial speech-to-text services (commercial speech-to-text services) trained on general speech data. Customer audio is processed in real-time and not retained for training by these providers under our Business Associate Agreements.

Human Oversight

Providers can review and correct transcripts before use. Transcripts are marked as AI-generated until reviewed.

Practice Knowledge Retrieval (RAG)

Purpose

Retrieves relevant protocols, standard operating procedures (SOPs), and clinical guidelines from practice-uploaded documents to assist with clinical decision-making.

Data Used

  • Practice-uploaded documents (protocols, SOPs, guidelines)
  • User query context
  • Session and patient context

Training Data

No external training. Uses practice-provided documents only. Retrieval is based on semantic similarity matching.

Human Oversight

Practice administrators control which documents are indexed. Retrieved information is presented as reference material, not clinical directives. Clinical judgment remains with the provider.

Denial Management & Appeal Assistance

Purpose

Analyzes insurance claim denials and suggests appeal strategies, generates draft appeal letters based on denial reasons and clinical documentation.

Data Used

  • Claim submission data and denial reasons
  • Clinical documentation supporting medical necessity
  • Payer-specific appeal requirements
  • Historical appeal outcomes (anonymized)

Training Data

Uses commercial AI models with Joyera-specific prompts. Appeal strategies are based on CMS guidelines, payer policies, and general healthcare billing best practices.

Human Oversight

All appeal letters require review by billing staff or providers before submission. Final appeal decisions rest with the practice.

Appointment Scheduling Optimization

Purpose

Suggests optimal appointment times based on provider availability, patient preferences, and historical scheduling patterns.

Data Used

  • Provider calendars and availability rules
  • Patient scheduling preferences
  • Appointment type and duration requirements
  • Historical no-show patterns (anonymized)

Training Data

Uses rule-based algorithms and commercial AI models. No patient-specific predictions about attendance are made.

Human Oversight

All scheduling suggestions require confirmation by staff or patients. No appointments are automatically booked without human approval.

Email Campaign Content Generation

Purpose

Generates draft content for patient communication campaigns (appointment reminders, health education, practice updates).

Data Used

  • Campaign purpose and target audience
  • Practice information and branding
  • General patient communication templates

Training Data

Uses commercial AI models trained on general marketing and healthcare communication best practices. No patient-specific data influences content generation.

Human Oversight

All campaign content must be reviewed and approved by practice staff before sending. Recipients can opt out of marketing communications.

AI Principles & Policies

Our AI Principles

Joyera's use of artificial intelligence is guided by principles of transparency, safety, and human oversight. We believe AI should enhance—not replace—the clinical judgment of healthcare professionals.

  • Assistive, Not Autonomous: All AI features are designed as assistive tools. No AI system makes clinical decisions, diagnoses patients, or takes actions without human review and approval.
  • Data Privacy: We do not use patient data (PHI) to train AI models. Our AI providers are bound by HIPAA Business Associate Agreements that prohibit using customer data for training.
  • Transparency: AI-generated content is clearly labeled. Users always know when they are interacting with AI-generated suggestions versus human-created content.
  • Human Override: Users can always edit, reject, or override AI suggestions. The AI serves the human, not the other way around.
  • Continuous Improvement: We monitor AI performance, collect user feedback, and regularly update our AI systems to improve accuracy and reduce errors.
  • Bias Mitigation: We actively work to identify and mitigate bias in AI outputs. Clinical suggestions are based on medical evidence, not demographic assumptions.

AI Providers and Data Processing

Joyera uses HIPAA-compliant AI providers, all of which have signed Business Associate Agreements. All AI processing occurs on servers located in the United States. No patient data is transferred internationally or used for model training.

Limitations and Risks

While we strive for accuracy, AI systems have inherent limitations:

  • Hallucinations: AI models may occasionally generate plausible-sounding but incorrect information. Always verify AI suggestions against clinical knowledge.
  • Incomplete Context: AI has access only to information provided in the current session. It may lack context from other sources.
  • Not Medical Advice: AI suggestions are not medical advice. Clinical decisions must be made by licensed professionals.
  • Evolving Technology: AI capabilities and limitations change over time. We regularly update our disclosures as technology evolves.

Your Rights

As a user of Joyera's AI features, you have the following rights:

  • Right to Explanation: You can request an explanation of how any AI feature works and what data it uses.
  • Right to Human Review: You can request human review of any AI-generated content or decision.
  • Right to Opt-Out: You can disable certain AI features while continuing to use Joyera's core services.
  • Right to Feedback: You can provide feedback on AI performance, which helps us improve our systems.
  • Right to Data Access: You can request a copy of data used by AI features in your account.

To exercise these rights, contact us at support@joyera.ai or submit a request via our privacy contact form.

NYC Department of Education Compliance

For school-based users subject to NYC Chancellor's Regulation A-820:

  • This AI Disclosure satisfies the transparency requirements for AI-powered education technology tools.
  • No student education records are used for AI model training.
  • AI features used for student services (IEP documentation, therapy notes) include mandatory human review by licensed professionals.
  • We maintain documentation of AI feature testing and validation for audit purposes.

ONC Health IT Compliance

Joyera is pursuing ONC Health IT Certification and is building toward compliance with Decision Support Intervention (DSI) requirements under 45 CFR 170.315(b)(11):

  • Source Attribution: All clinical decision support interventions include source attribution (e.g., FDA, CMS guidelines, clinical evidence).
  • Predictive DSI: When AI/ML models are used for predictive recommendations, we document risk analysis per NIST AI Risk Management Framework.
  • Feedback Mechanism: Users can provide feedback on AI interventions, which is logged and exportable for quality improvement.
  • Override Capability: All AI alerts and suggestions can be dismissed with documented reason codes.

Questions About Our AI?

We're committed to transparency. If you have questions about how we use AI, want to exercise your rights, or need additional information for compliance purposes, please contact us.

This disclosure is updated whenever we add or modify AI features. Check back regularly for the latest information.