Core Concepts

Peridot is built around a small set of core objects that define how AI is discovered, monitored, controlled, and audited across your environment.

Understanding these concepts is critical before configuring policies or integrations.

Workspaces

A workspace is the top-level container in Peridot.

It represents a logical boundary for your organization—typically aligned to a company, business unit, or security domain.

All inventory, data flows, policies, incidents, and integrations are scoped to a workspace.

Environments

Environments allow you to separate configurations within a workspace.

Common environments include:

Development
Staging
Production

Each environment can have its own policies, integrations, and controls.

AI Inventory

AI Inventory provides a complete view of AI tools used across your organization.

This includes:

External AI tools (e.g., ChatGPT, Claude)
Internal AI systems
Embedded AI features within SaaS platforms

Inventory distinguishes between:

Sanctioned tools (approved)
Shadow AI (unsanctioned or unknown usage)

Data Flows

Data Flows track how information moves into and out of AI systems.

This includes:

Source systems (files, SaaS, APIs, logs)
Destination systems (models, tools, workflows)
Context of usage

Data flows are the foundation for understanding risk and enforcing policies.

Applications

Applications represent AI-generated or AI-powered systems operating in your environment.

These may include:

Internal tools built with AI
Automated workflows
Embedded assistants

Peridot tracks these applications to ensure they are visible, governed, and auditable.

Policies

Policies define the rules that control AI behavior.

Examples include:

Which models can be used
What data can be accessed
Which integrations are allowed
When approvals are required

Policies are enforced across inventory, data flows, and applications.

Incidents

Incidents represent events that require investigation or response.

Examples include:

Sensitive data exposure
Unauthorized AI tool usage
Policy violations

Incidents can trigger workflows, alerts, and playbooks.

Integrations

Integrations connect Peridot to external systems.

These include:

SaaS platforms (Slack, Jira, ServiceNow)
Cloud providers (AWS, Azure, GCP)
Data stores and APIs

Integrations enable discovery, monitoring, and action.

Auditability

Every action in Peridot is traceable.

This includes:

Data access
Model usage
Policy decisions
Workflow execution

Auditability ensures that AI usage can be reviewed, explained, and defended.

What to do next

Read How Peridot Works to understand system architecture
Start with AI Inventory Overview to begin discovery