Skip to main content

Chatbot

/img/content-concepts-raw-nlp-chatbot-img.png

Introduction​

  • Definition: This product will automate tasks and handle conversations with the user.
  • Applications: Customer support, Product suggestion, Interactive FAQ, Form filling, Question Answering
  • Scope: Chat and Voice Support, FAQ, Knowledge-based and Contextual bot
  • Tools: DialogFlow, RASA, DeepPavlov, Alexa Skill, HuggingFace, ParlAI

Models​

RASA Chatbot​

RASA supports contextual conversational AI. It provided an integrated framework for Natural language understanding, dialogue generation and management. It also supprots multiple endpoints (e.g. Facebook messenger, WhatsApp) for easy deployment.

DialogFlow Chatbot​

It is an API to easily create and deploy chatbots. It also supports Voice interaction via Google cloud Voice API.

Alexa Skill​

This API enable us to create an alexa skill that can be used via alexa services. This also supports voice interaction via Alexa Voice API.

Process flow​

Step 1: Create As-Is Process

Create the current process

Step 2: Propose To-Be Process

Creeate the to-be process in which chatbot will handle the conversations in collaboration with human in the loop or on fully automated basis

Step 3: Collect the training data

Collect or create the training data and example conversations for training the chatbot

Step 6: Chatbot Training

Train the chatbot model

Step 9: UAT Testing

Wrap the model inference engine in API for client testing

Step 10: Deployment

Deploy the model on cloud or edge as per the requirement

Step 11: Documentation

Prepare the documentation and transfer all assets to the client

Use Cases​

RASA Chatbot​

Categorization of services and selected 4 most usable services for automation process. Development of a text-based chatbot application for this automation. RASA framework (python) was selected for implementation of this chatbot. Check out this notion.

Insurance Voicebot​

Automate the low-skill contact center services with the help of Voicebot AI technology. Context - Insurance Contact Centre, Role - A virtual customer care representative, Skills – Claim status, Language – English (US), Technology – Voice-enabled Goal-oriented Conversational AI Agents (Voicebots). Modules - Dialogflow Voicebot, Alexa Skill Voicebot, Rasa with 3rd-party Voice API, Rasa powered Alexa skill, Rasa powered Google assistant, Rasa voicebot with Mozilla tools, and DeepPavlov Voicebot. Check out this notion.

Wellness Tracker​

A bot that logs daily wellness data to a spreadsheet (using the Airtable API), to help the user keep track of their health goals. Connect the assistant to a messaging channelβ€”Twilioβ€”so users can talk to the assistant via text message and Whatsapp. Check out this notion.

RASA Chatbot Experiments​

Experiment with 4 chatbots in RASA: Financial Bot - A chatbot demonstrating how to build AI assistants for financial services and banking, Movie Bot - A bot to book movie tickets, Cricket Bot - A bot that will bring the live info about IPL cricket match as per user query, and Pokedex - This is a demonstration of a digital assistant that can answer questions about pokemon. Check out this notion.