1. Overview
  2. Getting Started
  3. Quick Start Guide

Quick Start Guide

Overview

This guide walks you through setting up Peridot and establishing initial control over AI usage in your environment.

By the end of this guide, you will:

  • Connect your first system

  • Discover AI usage across your environment

  • Monitor data flowing into AI systems

  • Apply your first governance policy

This process is designed to move from zero visibility to enforceable control in a single session.

Before You Begin

You will need:

  • Access to your Peridot workspace

  • Admin permissions

  • Access to at least one system (e.g. AWS, Slack, or a model provider)

  • An identity provider configured (optional but recommended)

Step 1 — Connect Your First System

Start by connecting a system where AI activity occurs.

Common starting points:

  • Cloud platform (AWS, Azure, GCP)

  • Communication tool (Slack)

  • Model provider (OpenAI, Anthropic)

This enables Peridot to begin discovering AI usage and collecting activity signals.

What Happens

  • Logs and events begin flowing into Peridot

  • AI-related activity is detected

  • Initial inventory starts to populate

Step 2 — Discover AI Usage

Navigate to AI Inventory.

Within minutes, you should see:

  • AI tools in use across your organization

  • Model providers being accessed

  • Early signals of AI-generated applications

What to Look For

  • Unknown or unsanctioned tools

  • Unexpected model usage

  • High-frequency AI activity

This is your first view into shadow AI.

Step 3 — Monitor Data Flows

Go to Data Flows.

Here you can observe:

  • What data is being sent to AI systems

  • Where that data originates

  • Which models are processing it

Why This Matters

Most risk comes from data exposure—not model usage itself.

Monitoring data flows gives you visibility into:

  • Sensitive data usage

  • External model interactions

  • Potential policy violations

Step 4 — Apply Your First Policy

Create a simple governance policy.

Example:

Policy: Restrict sensitive data from external models

  • Condition: Sensitive data detected

  • Rule: External models not allowed

  • Action: Reroute or block

What Happens

  • Future requests are evaluated in real time

  • Sensitive data is controlled automatically

  • Policy decisions are logged

Step 5 — Observe Enforcement

Once your policy is active:

  • Requests may be blocked or rerouted

  • Events are logged in real time

  • Incidents may be created

Go to Incidents Overview to review activity.

Step 6 — Review Audit Logs

Navigate to Audit Logs.

You will see:

  • Requests processed

  • Policy decisions

  • Enforcement actions

This provides a complete, traceable record of AI activity.

What You’ve Achieved

At this point, you have:

  • Discovered AI usage across your environment

  • Gained visibility into data flows

  • Applied governance policies

  • Observed enforcement in real time

  • Established auditability

You have moved from unmanaged AI usage to a controlled system.

Recommended Next Steps

  • Expand integrations (cloud, SaaS, identity)

  • Refine policies and routing rules

  • Configure incident playbooks

  • Set up SSO and role-based access

  • Integrate audit logs with your SIEM

In Production

In a production deployment:

  • All AI activity is monitored

  • Policies are enforced consistently

  • Data exposure is controlled

  • Incidents are managed systematically

  • Audit logs support compliance and investigation

Was this article helpful?