A STORY ON CONVERSATIONAL CHATBOT SUBHAJIT SENGUPTA & SHARIQ SHAIKH Can We Chat on Kanban A STORY ON CONVERSATIONAL CHATBOT BY SUBHAJIT SENGUPTA & SHARIQ SHAIKH
WHO ARE WE? WHAT WE DO? WHAT STORY? ©Digité
OUR JOURNEY WITH CHATBOT VISION EXECUTION Knowledge Base Dialogflow PLAN & DESIGN DEPLOYMENT Chatbot My Chatbot Wireframe Knowledge Base Google Web API Explore Platforms Application Explore Plug-ins Slack Skype ©Digité
“By 2020, the average person will have more conversations a day with bots than they do with their spouse” L. HEATHER PEMBERTON GARTNER ©Digité
CHATBOT TYPES THE HELPER THE SOLVER THE TEACHER THE SELLER ©Digité The Customer Service Chatbot like KLM’s chatbot The Utility Chatbot The Conversational Commerce Chatbot The Information Provider Chatbot ©Digité
<<Video/GIF on SwiftKanban Knowledge Base>> ©Digité
KNOWLEDGE BASE ©Digité ©Digité COMMENTS & FEEDBACK WORDPRESS BASED POWERFUL SEARCH FAQ AUTO-BUILD DYNAMIC TOC KNOWLEDGE BASE GOOGLE ANALYTICS DEVICE FRIENDLY CUSTOMER SUPPORT PERSONA BASED APPLICATION TASK ORIENTED ©Digité ©Digité
01 02 03 04 05 06 PoC ©Digité RESEARCH FRONTEND BACKEND STRUCTURE Explore Platforms & Plug-ins like Brainshop AI Chat Brainshop AI Chat My Chatbot My Chatbot 04 05 06 BACKEND STRUCTURE TESTING Training, Re-mapping DIalogflow Intent, Entities, Training Phrases, Integration ©Digité
WHY DIALOGFLOW No installation required, cloud based Easy to understand concepts, Intuitive UI Fast prototyping of chatbots, inbuilt 'try-it' Very less training time Integration with all major platforms SmallTalk for common conversations Powered by Google's strong NLU algorithms ©Digité
BACKBONE OF DIALOGFLOW AGENT : Natural Language Understanding (NLU) modules INTENT: Maps user input to responses ENTITY: Identifies and extracts useful data from natural language inputs CONTEXT: Represents the current state of a user's request and allows your agent to carry information from one intent to another FULFILLMENT: Fulfillment is code that's deployed as a webhook that lets your Dialogflow agent call business logic on intent-by-intent basis ©Digité
©Digité
AGENT A module which can be included in your app, website, product, or service and translate text or spoken user requests into actionable data. This translation occurs when a user's utterance matches an intent within your agent. ©Digité
BACKEND INTERFACES WordPress Plugin Dialogflow ©Digité ©Digité
ENTITY System Entity Custom Entity Custom Entity Entities pick out specific pieces of information that users mention. System Entity Custom Entity Custom Entity @sys.date-time @sys.unit-area @sys.address @Kanaban.card @Kanban.board @Team_Dynamics ©Digité
©Digité
INTENT In each intent, you define examples of user utterances that can trigger the intent, what to extract from the utterance, and how to respond. Training Phrase Actions & Parameters Response ©Digité
Matched Utterances Matched Entities Response ©Digité
©Digité
DEFAULT FALL BACK INTENT I do not know the answer! What do I do? Respond to user if there's no matching intent for the query How can we use fallback intent for better User Experience? The idea of Google Web Search API ©Digité
SMALL TALK How are you! Custom Entity I’m so excited! Custom Entity You're beautiful I’m so excited! Custom Entity You're beautiful You’re beautiful! You're beautiful Who are you! ©Digité
INTEGRATION Integration with Google Assistant, Slack, Facebook Messenger, Twitter. Skype, Skype for Business, Amazon Alexa, Microsoft Cortana and so on. SUBHAJIT ©Digité
ANALYTICS Usage data: Number of sessions and queries per session. NLU data: Most frequently used intents and exit percentages. ©Digité
ROADMAP 01 02 03 04 Slack & Skype Actionable Test Automation RASA SUBHAJIT Slack & Skype 02 Actionable 03 Test Automation RASA 04 ©Digité
https://dialogflow.com/docs READ https://dialogflow.com/docs SUBHAJIT WATCH https://goo.gl/bmdwTc CONTACT subhajit1975@gmail.com / shariq.shaikh191@gmail.com ©Digité
THANK YOU