Built Chatbot On This Blog With Prince Jain

Harshal

Dec 3, 20242 min read

V1 Of Making Content More Accessible, Engaging, And Interconnected

When coaching or consulting clients, I frequently refer to my blog to provide detailed, visual answers. I wanted to avoid repeating myself for FAQs, but every question was phrased differently. I also wanted to ensure that new posts interlink with existing content to make it easy for users to navigate.

To address this, I collaborated with Prince Jain to create a solution using GenAI. Prince developed a chatbot and launched it on the home page. You can find Prince's explanation and use his chatbot as a template here.

I spent 40 minutes defining this chatbot and 24 minutes writing this blog. You need 2 minutes to read this.

I first published this on www.sparkcreativetechnologies.com on Dec 03, 2024.

I tested it by asking a sample question about building strong relationships. The chatbot pulled information from the relevant blog post.

Sample user experience: ask a question, get a response.

Chatbot Project Goals

I defined the following requirements for the chatbot. Think of this as the Product requirements document.

1 - Blog Content Indexing and Searchability

Index all website content, focusing on harshal-patil.com.
Refresh the index weekly to accommodate frequent updates.

2 - Original Content Attribution

Attribute content to sparkcreativetechnologies.com or harshalpatil.substack.com where required, depending on the source (not just the harshal-patil website).

3 - Content Searchability

Enable users to search past blog posts and receive direct links.
Provide synthesized answers from blog content for user queries.

4 - Blog Post Interlinking

Suggest interlinking opportunities for new posts to boost traffic across the blog.
Identify related old blogs and suggest specific sections for adding links. Example: When I write a new post about revenue growth through customer journey mapping, the tool should recommend linking to relevant sections in older posts related to understanding customers or growth.

How The Developer Built The Chatbot

Building a chatbot from scratch offers unmatched flexibility and scalability compared to pre-built plugins. Here is Prince’s guide to creating a customizable chatbot framework.

Tools You’ll Use:

Python: Handles backend logic, APIs, and AI response processing.
OpenAI (or any LLM): Processes user queries.
Google Cloud: Ensures reliable and scalable backend hosting.
Wix: Integrates the chatbot into your website with a user-friendly interface.

Steps to Build the Chatbot:

Clone the code.

git clone https://github.com/Princejain1101/pythonchatapp

Set up a Python virtual environment and install dependencies.
Configure the OpenAI API key and test locally.

python main.py

Deploy to Google Cloud.

gcloud app deploy

Add a Wix chatbot and link it to your backend using custom JavaScript.

Try The Chatbot

You can test the chatbot here.

Next Steps

On launching this Chatbot, I realized some gaps in the user experience and defined a new version. I’ll share examples and the new user experience, with examples of what the new version will accomplish.