Agent Registry

What is the Agent Registry?

The Agent Registry is a catalog of the Appian AI agents you can access in your environment. The registry lets you discover and oversee agents from one location. Every Appian AI agent you have permission to see appears automatically.

Why use the Agent Registry?

Use the Agent Registry to understand which agents exist, who owns them, and how they are performing. The registry provides the following advantages:

• Find existing agents to reuse: Find an existing agent that fits your use case more quickly.
• Visibility: See ownership and 30-day health for every agent you can access from one screen.
• Cost awareness: See how many AI actions each agent consumes.

Access the Agent Registry

You open the Agent Registry from the Agent Controls view in Appian Designer. Which agents display depends on where you open it from:

• Environment level: The registry lists every agent you can access across all applications.
• Application level: The registry automatically narrows to that application's agents.

Find and view agents

In the Agent Registry, you can switch between a grid and list view, which displays each agent's:

• Name
• Owner
• Application
• Custom tools
• 30-day metrics

To find a specific agent, use the search, filter, and sort options.

Scroll or drag to pan · ⌘+/⌘− to zoom · Esc to close

Select an agent's name to open a detail view, which includes the agent's:

• Description
• Details (owner, application, model, and agent type)
• 30-day performance metrics
• Custom tools
• AI guardrails enabled in the environment.

Click OPEN IN DESIGNER to edit the AI agent.

Scroll or drag to pan · ⌘+/⌘− to zoom · Esc to close

Agent metrics

The Agent Registry reports each agent's operational metrics over a fixed, rolling 30-day window.

The grid view shows the following metrics for each agent:

• Total Agent Runs: The number of runs the agent completed in the last 30 days.
• Total AI Actions: The AI actions the agent consumed in the last 30 days.
• Avg Latency: The average duration of the agent's successful runs, calculated over the last 30 days. Only applies to process agents; chat agents show N/A.
• Success Rate: The percentage of the agent's runs that completed without an error in the last 30 days. Success rate measures error-free runs, not whether the agent's output was correct.

The detail view adds two AI action metrics:

• Avg AI Actions/Run (last 30 days): The average AI actions per run, calculated over the last 30 days.
• AI Actions/Run (Today): The average AI actions per run for the current day, so you can compare recent activity against the 30-day average.

Scroll or drag to pan · ⌘+/⌘− to zoom · Esc to close

For the full history of an individual agent's runs, including the tool calls, inputs, and outputs for each run, go to the Monitor tab.

Custom tools

Custom tools that display in the Agent Registry include the names of the custom tools and external tools that are added to the agent so you can see what each agent does at a glance.

The following are listed under custom tools:

• Process models: The name of each process model the agent can start.
• Expression rules: The name of each expression rule the agent can call.
• AI agents: The name of other agents the agent can call.
• MCP tools: The name of each Model Context Protocol (MCP) tool the agent can use.

Who can see agents in the registry

The Agent Registry respects object security. You see only the agents you have at least Viewer permission to.

Feedback