Upgrade to Pro — share decks privately, control downloads, hide ads and more …

Creating Conversations For Chatbots and Voice Activated speakers

Creating Conversations For Chatbots and Voice Activated speakers

Lee Boonstra

June 01, 2018
Tweet

More Decks by Lee Boonstra

Other Decks in Technology

Transcript

  1. 8

  2. 12 Chatbots exists since the 90’s… So why are they

    popular now? 1994: First Chatbot 2016: Focus on AI First
  3. 13 70% of requests are in natural language Is it

    gonna rain tomorrow? Tomorrow it will be sunny In Amsterdam!
  4. 15 It’s easier to make computers learn than to build

    smarter computers Machine Learning Process of making computers learn Artificial Intelligence Process of building smarter computers
  5. Confidential & Proprietary Google Home — Voice-activated speaker powered Google

    Assistant — A conversation between you and Google that helps you get more done in your world
  6. Confidential & Proprietary Example of a Custom Use Case Ok

    Google, let me talk to BookAMeetingRoom Welcome, how can I help you? I want to book a meeting room for 3 persons. Sure! Here’s BookAMeetingRoom Sure, for when? Tomorrow, from 2pm to 3pm.
  7. 25 • End-to-end development suite for building conversational UI. •

    Previously known as API.AI • Sept 2016, acquired by Google • Powered by Machine Learning Dialogflow
  8. 26 Why Dialogflow • Powered by Machine Learning (NLP) •

    Scalable: separate your conversation text from code • Build faster with the Web UI • Advanced fulfillment options and multi channel integrations • Analytics • Speech / Voice Integration • Multi-lingual bot support (20+ languages) • Part of Google Cloud (60+ cloud services) Intent Recognition (NLP) ... Web Console Powerful SDK Supervised Training Cross-platfor m
  9. 28 3 demo’s for the following user profiles: CONTENT WRITER

    • Dialogflow Web UI (intents, entities, training, analytics, integration) ENGINEER • Fulfillment (webhooks) • Front-end END USER • Multi Channel (Web, Google Assistant) Dialogflow Demo
  10. 29 • Entities vs. System Entities • Intents • Training

    • Analytics • Webhooks • Integration ◦ Example: For Web ▪ Welcome Message ▪ Rich Cards ◦ Example: For Google Assistant Dialogflow Demo
  11. 30 Multi Channel Build once. Integrate everywhere. Dialogflow Enterprise Content

    Writer, creates Conversational UI in Dialogflow User communicates With Agent. Deploy to websites & apps Deploy to chat channels and social media Deploy to voice activated speakers and assistants Deploy for phone and SMS services Dialogflow replies to a user session
  12. 31 Dialogflow build process Setup GCP Account Design your Conversation

    in Dialogflow Enterprise Deploy for multi-channel Integrate with internal systems & APIs SDK, GCP integrations Test your agent Agent Training Agent Analytics Tweak your conversation in Dialogflow Enterprise Optional: Engineer Optional: Engineer Business / IT Content Writer Content Writer Content W riter
  13. 32 Technical Designs Human bot take over User types in

    chat UI Chatbot or User replies Operator Dashboard Kubernetes Engine Container Registry Customer Client Kubernetes Engine Back-end Kubernetes Engine Back-end Kubernetes Engine NLP API ML call for sentiment detection Pass user utterance to chatbot Return text response to user User asks to talk with human Negative sentiment let a human take over the conversation Operator types in chat UI User replies Containers images can be stored in the Container Registry Dialogflow Enterprise
  14. Confidential + Proprietary Advanced Chatflow with machine learning bot analytics

    User types to custom UI or channel Chatbot replies Dialogflow Enterprise Customer Client JS Angular 5 web front-end Kubernetes Engine Chat Server Dialogflow SDK / socket.io Kubernetes Engine Back-end CRM Python / Django Kubernetes Engine Container Registry Containers images can be stored in the Container Registry Messaging Publisher Pub/Sub Webhook Router Cloud Function Webhook Container Builder Building Dev Pipelines
  15. Confidential + Proprietary Advanced Chatflow with machine learning and bot

    analytics User types to custom UI or channel Chatbot replies Dialogflow Enterprise Customer Client JS Angular 5 web front-end Kubernetes Engine Chat Server Dialogflow SDK / socket.io Kubernetes Engine Back-end CRM Python / Django Kubernetes Engine Subscription Cloud Function Sensitivity Filter DLP API Sentiment Detector NLP API Data Warehouse BigQuery Messaging Publisher Pub/Sub Webhook Router Cloud Function Webhook
  16. 35 Cloud Functions Completely serverless & fully managed service ⇒

    don’t worry about the ops! Automatic scaling and super-fast ⇒ grows with the success of your project ⇒ cost-effective Open and familiar ⇒ JavaScript / Node.js
  17. Confidential & Proprietary Review & Approval Web-based portal • Triggering

    Information • Merchandising and information Approvals • Automatic and manual policy checks • Turn around in about 1 week
  18. Confidential & Proprietary Demo: A TV Guide Do it yourself:

    https://goo.gl/pH86aT Ok Google, let me talk to the TV Guide Welcome, I’m the Dutch TV Guide. You can ask me the following: What’s on a channel? What’s on RTL 4? Sure! Here’s The TV Guide Enter Earcon On RTL 4: At 7am: Nieuws At 7:30am: Nieuws ... Exit Earcon . . .
  19. Confidential & Proprietary Resources • Google Dev Docs - https://developers.google.com/actions/develop/conversation

    • Examples - https://github.com/actions-on-google • Dialogflow - https://console.dialogflow.com • Simulator - https://developers.google.com/actions/tools/web-simulator • Hands-on Workshop Materials - https://goo.gl/pH86aT