Resources

Definitions

Key terms across governed execution — healthcare AI and the software supply chain

Contact us

Glossary

A–D

API — Application Programming Interface

A set of protocols and tools that lets different software systems communicate. In healthcare, APIs enable AI and scheduling systems to exchange data securely.

Agentic AI

AI systems that can plan, take actions, and use tools to accomplish goals—such as searching for providers or booking appointments—rather than only answering questions.

Attestation

Cryptographic proof, issued by a hardware security token, that a key was generated inside the hardware and can never be exported. SignetKeys uses attestation to prove a signing key never existed as a file.

Code Signing

Cryptographically signing software so operating systems can verify who published it and that it hasn’t been tampered with. On macOS, unsigned or badly signed software is blocked by Gatekeeper.

CSR — Certificate Signing Request

A file containing a public key and identity details, submitted to a certificate authority (such as Apple) to obtain a signing certificate. The private key it corresponds to never needs to leave the hardware that generated it.

Developer ID

Apple’s certificate type for distributing macOS software outside the App Store. Developer ID–signed and notarized software is what Gatekeeper allows to run without warnings.

Embedding

A numerical representation of text or data that captures meaning. Embeddings let AI find semantically similar content—e.g., matching patient intent to the right provider or service.

F–H

FHIR — Fast Healthcare Interoperability Resources

A modern standard for exchanging healthcare data. FHIR makes it easier for AI and scheduling systems to access appointments, providers, and patient information across platforms.

Gatekeeper

The macOS security system that checks whether software is signed by a valid Developer ID and notarized by Apple before allowing it to run. The reason code signing is mandatory for macOS distribution.

Governance

Policies and controls that define what an automated system can do, on whose behalf, and how every decision is logged and reviewed. The common thread across PersonixHealth products.

Hardware Custody

Holding a cryptographic key inside a physical security token that performs operations on the key’s behalf but can never reveal it. Contrast with keys stored as files, which can be copied and stolen.

Hallucination

When an AI model generates plausible-sounding but incorrect or fabricated information. Mitigated by grounding responses in trusted data (e.g., via RAG) and deterministic orchestration.

HIPAA — Health Insurance Portability and Accountability Act

U.S. law that protects the privacy and security of health information. Any AI handling patient data must comply with HIPAA’s requirements for access, storage, and disclosure.

HL7

Health Level Seven—a standards organization that develops frameworks for healthcare data exchange. FHIR is an HL7 standard widely used for interoperability.

L–O

LLM — Large Language Model

AI models trained on vast amounts of text to understand and generate human language. Used for conversational interfaces, summarization, and intent understanding in healthcare.

MCP — Model Context Protocol

A protocol that lets AI models connect to external tools and data sources in a standardized way. Enables secure, governed access to enterprise capabilities like scheduling and search.

NLP — Natural Language Processing

The branch of AI that enables computers to understand and process human language. Powers conversational interfaces, intent detection, and search in patient engagement tools.

Notarization

Apple’s automated malware scan of signed macOS software. Accepted submissions receive a ticket that is “stapled” to the software so Gatekeeper can verify it even offline. Required for smooth distribution outside the App Store.

OAuth

An authorization framework that lets users grant applications limited access to their data without sharing passwords. Used for secure, delegated access in healthcare integrations.

OIDC — OpenID Connect

An identity layer on top of OAuth that lets a system prove who it is with short-lived, verifiable tokens instead of stored secrets. How SignetKeys authenticates CI pipelines with zero credentials in the build.

P–Z

PHI — Protected Health Information

Individually identifiable health information held by covered entities. Must be safeguarded under HIPAA; AI systems must handle PHI with appropriate security and consent.

PII — Personally Identifiable Information

Data that can identify a specific person—name, email, phone, etc. In healthcare, PII often overlaps with PHI and requires careful handling and privacy controls.

Prompt

The instructions or input given to an AI model to guide its response. Well-designed prompts help ensure accurate, relevant, and safe outputs in healthcare applications.

RAG — Retrieval-Augmented Generation

A technique where an AI model retrieves relevant documents or data before generating a response. Reduces hallucination by grounding answers in trusted enterprise knowledge.

Software Supply Chain

Everything that goes into building and delivering software: dependencies, build systems, CI pipelines, and signing. A prime attack surface—compromising one build step can compromise every downstream user.

SSO — Single Sign-On

Authentication that lets users log in once to access multiple applications. Simplifies access for staff and supports secure integration of AI tools into existing workflows.

Vector Database

A database optimized for storing and querying embeddings. Enables fast semantic search—finding content by meaning rather than exact keywords—for AI-powered healthcare applications.

Have questions about these terms?

Talk to PersonixHealth