Functions

Here’s a complete breakdown of how to configure and set up the Functions Skill in raiaAI’s Launch Pad. This skill allows your AI agents to execute backend functions—such as retrieving CRM data, booking appointments, triggering actions in other software, or querying databases—based on user intent.

---

🧩 Functions Skill Overview

The Functions Skill empowers your agent to:

• Run custom logic like get_lead_status, schedule_meeting, fetch_order_status

• Use real-time data from external sources

• Interact with APIs, internal databases, or third-party services

This brings dynamic capabilities to your agent beyond static knowledge or basic Q&A.

---

🛠️ Functions Skill Setup Guide

---

1. Enable the Skill

• Navigate to your agent's Skills tab

• Click “Add Skill”

• Choose Functions from the list and toggle it Active

---

2. Add a New Function

Go to the Functions Library (can be global or agent-specific):

• Click “Create Function”

• Define the following:

🧾 Function Definition Fields

Field	Description
Name	Identifier for the function (e.g., get_user_profile)
Description	What the function does—shown to the AI agent to decide when to use it
Inputs	Define expected parameters (e.g., email, user_id, date)
Outputs	Describe expected response data
Endpoint URL	The webhook/API endpoint the function hits
Method	Choose GET, POST, PUT, etc.
Headers	Include any required headers (e.g., Authorization keys)
Body Template	If POST, define dynamic JSON payload using input variables
Test Button	Use mock values to verify the function runs correctly

You can also add logic for:

• Success/Failure Handling

• Fallback responses

• User-friendly error messages

---

3. Function Usage in Agent Conversations

Once the function is created:

• The AI agent can automatically decide when to call it based on user prompts

• Example: User: "Can you check the status of my order?" → Agent: (detects intent) → calls get_order_status function → returns result to user

---

4. Testing & Validation

• Use Copilot to simulate conversations

• Verify the function is:

  • Triggered under the right conditions

  • Receiving valid input/output

  • Returning readable and useful results to users

---

5. Organize & Reuse

• Save commonly used functions in the Global Functions Library

• Functions can be reused across multiple agents

---

✅ Best Practices

Tip	Why It Helps
Use clear descriptions	Helps the LLM know when to trigger the function
Include default values in test cases	Speeds up validation
Keep response outputs structured	Easier for agent to parse and summarize
Sanitize inputs & outputs	Ensure safe and user-friendly interactions
Monitor usage	Track which functions are called and how often (for improvements)