Skip to main content

Enable your AI Agent to take action with Agent Tools

Build and connect powerful tools your AI Agent can use to fetch data, perform actions, or trigger workflows

Alvaro Vargas avatar
Written by Alvaro Vargas
Updated over a week ago

Agent Tools allow you to give your AI Agent superpowers — by connecting it to external systems via APIs. Whether it’s checking an order status, pulling transaction history, or sending a webhook, your Agent can now make live API calls mid-conversation — without any coding required in the flow.

🛠 What are Agent Tools?

Agent Tools are custom-built API integrations that you can create and manage directly in Frontline. Once created, these tools can be assigned to any AI Agent node in your flow.

When the Agent determines that a tool is relevant (based on the user’s message and your instructions), it will invoke the tool automatically, retrieve the result, and use it to respond intelligently.

✏️ How to Create a Tool

  1. Go to Tools in the main sidebar

  2. Click “Create Tool”

  3. Fill in:

    1. A tool name your agent can easily recognize (Get Order Status)

    2. A description explaining when to use it (e.g. “Use when the user asks about their order status”)

  4. Connect your API:

    1. Supported methods: GET, POST, PUT, PATCH, DELETE

    2. Add any data inputs (like order_id, user_email, etc.)

    3. You can interpolate values into the URL using curly braces, e.g.: https://api.myshop.com/orders/{order_id}

    4. Add query parameters, headers, and test your request

⚙️ Assigning Tools to the AI Agent Node

To use a tool in a flow:

  • Drag in an AI Agent node

  • Go to the Tools tab inside the node

  • Select the tool(s) you want this agent to have access to

During the conversation, if the AI Agent determines a tool is needed, it will call it.

🔍 Tool Usage in Conversation Audits

Every time a tool is used, you can review:

  • Which tool was called

  • What data was sent

  • The API response

  • The final AI reply

Just open the Conversation Audit to trace every step with full transparency.

🔄 Example: Order Status Tool

Let’s say your users often ask, “Where’s my order?”

  1. Create a tool called Order Status

  2. Set up the API call with order_id as a dynamic input

  3. In the “When to use” field, write:

    1. Use this tool to check the status of an order based on the order ID

  4. Assign it to your Agent in a flow

  5. When a user asks for an update, the Agent will:

  6. Capture the order ID

  7. Call the tool

  8. Reply with real-time shipping info

Did this answer your question?