Policy-Based Routing

Overview

Policy-based routing controls how AI requests are directed across models and providers.

Instead of hardcoding model selection, Peridot routes requests dynamically based on policies.

Why Routing Matters

Without routing:

  • Developers choose models manually

  • Sensitive data may be sent to the wrong provider

  • Governance is inconsistent

With routing:

  • Model selection is centralized

  • Policies enforce compliance automatically

  • Requests are routed safely and consistently

How Routing Works

When a request is made:

  1. The request is evaluated against active policies

  2. Conditions determine applicable rules

  3. The system selects an approved model or provider

  4. The request is routed accordingly

Common Routing Patterns

Sensitive Data Routing

  • Route sensitive data to internal or approved models

Cost Optimization

  • Route low-risk requests to lower-cost models

Environment-Based Routing

  • Dev → sandbox models

  • Prod → approved production models

Multi-Model Workflows

Peridot supports routing across multiple steps:

  • Retrieval → internal system

  • Reasoning → approved model

  • Execution → API or integration

Each step can be governed independently.

Benefits

  • Centralized model control

  • Reduced risk of data exposure

  • Consistent behavior across applications

  • Flexibility without code changes

Next Steps

  • Learn how to [Create Policies]

  • Configure [Enforcement Actions]

  • Review supported model providers


Was this article helpful?