Enhancing the banking experience for both English and Yoruba speakers with seamless, intuitive support.
# Bi-Lingual BankerBOT Chatbot
## 🚀 Introduction
Building a chatbot that truly connects with people isn’t just about code—it’s about understanding their needs. That’s exactly what I set out to do with Bi-Lingual BankerBOT. This chatbot helps users navigate their banking needs effortlessly, supporting both English and African languages for a smoother and more inclusive experience.
## 📌 Table of Contents
- About Amazon Lex
- Project Setup
- Key Features
- Tech Stack
- Epic User Story
- Use Cases
- Intents Configuration
- Fallback Mechanism
- Testing and Iteration
- Conclusion
## 🤖 About Amazon Lex
Amazon Lex is the engine that powers BankerBOT. It’s a cloud-based AI service that makes it incredibly easy to build chatbots. With its built-in speech recognition and natural language processing, I was able to focus on crafting the best user experience instead of wrestling with complex AI models.
## ⚙️ Project Setup
**Amazon Lex Setup**
### Step 1: Creating the Chatbot
Getting BankerBOT up and running took just 3 minutes. The interface is intuitive, making it simple to create a functional chatbot without breaking a sweat.
### Step 2: Assigning Permissions
Security is key, so I created a role with basic permissions to ensure the bot has just enough access to function—nothing more, nothing less.
### Step 3: Setting Confidence Threshold
I stuck with the default confidence threshold of 0.40, which helps the bot balance between accurate responses and reducing false positives. This ensures smoother interactions and fewer frustrating moments for users.
## 🌟 Key Features
- **🗣️ Bilingual Support**: Handles conversations in both English and African languages.
- **🔍 Smart Intent Recognition**: Accurately understands user queries and provides meaningful responses.
- **🛡️ Robust Fallback System**: Ensures that even unrecognized inputs receive helpful guidance.
- **💬 Natural Conversation Flow**: Uses multiple variations to make interactions feel human-like.
## 🛠️ Tech Stack …