Function Creation and Webhook Setup Guide

Introduction

This guide explains how to create functions for your Peki chatbot and set up a webhook server to handle data processing. Functions allow your chatbot to collect specific information from users and send this data to your webhook server for processing.

Dashboard Access

The Peki Dashboard is accessible at:

https://dashboard.peki.ai/dashboard/{chatbotId}/chatbot/functions

Important: Replace {chatbotId} with your actual chatbot ID from the Peki platform.

Process Overview

1. First, we'll create a function using the Peki Dashboard
2. Then, we'll set up a webhook server to process the data
3. Finally, we'll test the integration

When a user interacts with your chatbot:

• The AI collects required information using defined properties
• The data is sent to your webhook server using the HTTP Action
• Your server processes the data and sends back a response
• The AI presents the results to the user using the response template

Function Configuration Fields

1. Basic Information

Field	Example Value	Description
Name	Credit Limit Check	Display name of the function
Description	Important: Collects necessary parameters from user to check credit score and determine limit.	Detailed explanation of function's purpose

2. Properties Configuration

Each property requires the following fields:

Field	Type	Required	Description
Name	String	Yes	Unique identifier for the property
Description	String	Yes	Instructions for AI to collect the information
IsRequired	Boolean	Yes	Whether this property must be collected

Example Properties Table

Name	Description	IsRequired
bank	Name of the bank to check credit score	Yes
loanAmount	Amount of loan to be taken	Yes

3. HTTP Action Configuration

Field	Example Value	Description
Name	Http Action	Display name of the action
Url	http://localhost:3001/webhook/credit	Endpoint URL for the webhook
Method	POST	HTTP method (GET or POST)

4. Response Configuration

Field	Example Value	Description
Response	Summarize the response from {bank} for loan amount {loanAmount}	Template for AI to format response

Function Creation Process

1. Access Dashboard

• Navigate to the Peki Dashboard using your chatbot ID
• Click "Create New Function"

2. Configure Basic Settings

Field to Fill	What to Enter
Name	Your function name (e.g., "Credit Limit Check")
Description	Clear explanation of function purpose

3. Add Properties

Step	Action
1	Click "Add Property"
2	Enter property name (e.g., "bank")
3	Write clear description
4	Set Required flag
5	Repeat for all properties

4. Configure HTTP Action

Step	Action
1	Enter webhook URL
2	Select HTTP method
3	Save action settings

5. Set Response Template

Step	Action
1	Write response template using {propertyName} syntax
2	Test template with sample values

Webhook Setup

1. Basic Node.js Server Setup

const express = require("express");
const bodyParser = require("body-parser");

const app = express();
app.use(bodyParser.json());

2. Create Webhook Endpoint

app.post("/webhook/credit", async (req, res) => {
  try {
    const { attributes } = req.body;

    // Extract user and function data
    const {
      chatUserId, // Unique chat user ID
      platformId, // User ID from script
      bank, // Collected by AI
      loanAmount, // Collected by AI
    } = attributes;

    // Process credit check logic
    const creditResult = await processCreditCheck({
      userId: platformId,
      bank,
      loanAmount,
    });

    // Send response
    res.json({
      success: true,
      data: creditResult,
    });
  } catch (error) {
    console.error("Webhook error:", error);
    res.status(500).json({
      success: false,
      message: error.message,
    });
  }
});

3. Start Server

const PORT = process.env.PORT || 3001;
app.listen(PORT, () => {
  console.log(`Webhook server running on port ${PORT}`);
});

Function Execution Flow

Step	Description
1. AI Trigger	AI recognizes intent and starts property collection
2. Data Collection	AI asks user for each required property
3. Validation	System ensures all required properties are collected
4. Webhook Call	System sends data to webhook URL
5. Processing	Webhook processes request and returns result
6. Response	AI formats response using template and presents to user

Best Practices

1. Property Naming

Do	Don't
Use clear, descriptive names	Use generic names
Use camelCase or snake_case	Use spaces or special characters
Keep names concise	Use very long names

2. Property Descriptions

Do	Don't
Write clear instructions	Use technical jargon
Include format requirements	Be vague or unclear
Specify any constraints	Assume AI understands context

3. Webhook Implementation

Practice	Description
Security	Use HTTPS in production
Validation	Check all incoming data
Error Handling	Implement proper error handling
Logging	Log all requests and responses
Timeouts	Include timeout handling

4. Testing

Test Case	What to Check
Happy Path	Basic successful flow
Missing Data	Handling of missing properties
Invalid Data	Error handling for invalid input
Response Format	Correct template processing

Troubleshooting

Issue	Solution
Webhook not receiving data	Check URL and network settings
Missing properties	Verify property configuration
Response not formatting	Check template syntax
Server errors	Check logs and error handling