Automate WhatsApp Conversations with Low-Code: A Guide using chatGPT API

Table of Contents

1. Introduction
2. The Need for AI Assisted Jira Ticket Recording
3. Building the WhatsApp Webhook
   1. Verifying the Webhook
   2. Setting up an HTTP Proxy with Ngrok
4. Capturing and Saving Voice Messages
   1. Downloading Audio Messages
   2. Converting Audio to Wave Format
5. Speech Recognition
   1. Utilizing Azure Speech Recognizer
6. Generating Well-Formatted Tickets
   1. Using the AI Ticket Generator
7. Inserting Issues into Jira
   1. Creating Jira Issues
8. Conclusion
9. FAQs

Introduction

In this article, we will explore how to build an AI-assisted Jira ticket recorder using CRANQ, a low-code development platform. We will utilize various APIs and services, such as WhatsApp Business Cloud, Azure Cognitive Services, and Atlassian API, to Create a seamless workflow for recording tickets using voice messages. This tutorial aims to provide developers with practical skills in building applications that leverage AI and improve productivity in issue tracking systems like Jira.

The Need for AI Assisted Jira Ticket Recording

Issue tracking is a crucial aspect of software development, particularly when working in a team. However, manually creating and updating tickets in Jira can be time-consuming and tedious for developers. As developers, we often wish for a more efficient way to communicate our tasks to Jira, especially when we have multiple tickets to create and update.

Building the WhatsApp Webhook

Verifying the Webhook

Before we can start processing incoming messages from WhatsApp, we need to verify our webhook. WhatsApp sends a GET request with a challenge STRING to ensure our webhook is ready to receive messages. We will set up an HTTP proxy using Ngrok to make our locally developed webhook accessible to WhatsApp.

Setting up an HTTP Proxy with Ngrok

Ngrok is a cross-platform tool that allows us to create a secure tunnel and forward HTTP traffic from our local development environment to the outside world. By using Ngrok, we can expose our webhook on port 8080 and obtain an activated URL domain that WhatsApp can use to send requests to our webhook.

Capturing and Saving Voice Messages

To enable voice-Based ticket recording, we need to capture and save voice messages sent through WhatsApp. We will download the audio messages using WhatsApp's media downloader and then convert them to wave PCM format, which is required by Azure Speech Recognizer.

Downloading Audio Messages

WhatsApp provides a media ID for each audio message, which we can extract from the webhook request. By utilizing the WhatsApp media downloader, we will save the audio messages to our local file system.

Converting Audio to Wave Format

Azure Speech Recognizer expects audio files in wave PCM format. However, the audio messages we downloaded from WhatsApp are in .ogg format. To satisfy the format requirement, we will use the FFM pack-based Wave Converter node in CRANQ to convert the .ogg files to wave format.

Speech Recognition

Once we have the audio messages in wave format, we can utilize Azure Speech Recognizer to convert the speech into text. The Azure Speech Recognizer node takes the converted wave file and uses Azure Cognitive Services to transcribe the speech. We will extract the recognized text for further processing.

Generating Well-Formatted Tickets

To create well-formatted tickets, we will leverage the power of open AI's ChatGPT system. We will use the AI Ticket Generator node, which interprets free-worded text descriptions as user stories. This node breaks down the text into summary, description, acceptance criteria, and priority fields required for a Jira ticket.

Inserting Issues into Jira

To complete the ticket recording process, we will insert the generated issues into Jira using the Jira Issue Creator node. This node interacts with the Atlassian API and requires authentication credentials such as email, API token, Jira domain, project key, and issue Type. We will adapt these credentials from the environment variables set earlier.

Conclusion

In this tutorial, we have learned how to build an AI-assisted Jira ticket recorder that allows developers to create and update tickets using voice messages sent through WhatsApp. By integrating various APIs and services, we have streamlined the process of capturing voice messages, transcribing speech, and generating well-formatted issues. This low-code approach not only saves time but also provides developers with practical skills in leveraging AI to enhance their development workflows.

FAQs

Q: Can I use this ticket recorder with any issue tracking system other than Jira? A: The tutorial focuses on integrating with Jira using the Atlassian API. However, with some modifications, you can adapt the solution to work with other issue tracking systems that provide APIs for ticket creation and management.

Q: Is it possible to add additional fields to the Jira ticket template? A: Yes, the ticket template used in this tutorial includes summary, description, acceptance criteria, and priority fields. You can customize the template and the corresponding mapping in the Jira Issue Creator node to include additional fields specific to your project requirements.

Q: Can I use a different speech recognition service instead of Azure Speech Recognizer? A: Yes, you can replace the Azure Speech Recognizer node with other speech recognition services that provide APIs. However, you will need to modify the integration accordingly to adapt to the API and authentication requirements of the chosen service.

Q: How secure is the communication between WhatsApp and CRANQ in this setup? A: The communication between WhatsApp and CRANQ is secured using HTTPS and can be further enhanced by utilizing NGROK's secure tunnel. However, it is essential to follow best practices for securing your local development environment and handling sensitive data, such as authentication tokens and API keys.

Q: Are there any limitations or performance considerations when using the free tiers of the services Mentioned? A: It is important to be aware that free tiers of services may have certain limitations, such as rate limits, usage quotas, or reduced functionality. Before deploying the solution to a production environment, consider the requirements and limitations of the specific services you are using and ensure they meet your needs.