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

Chatbots, et si on passait la seconde ?

Chatbots, et si on passait la seconde ?

Workshop on implementing chatbots for the Google Assistant using Dialogflow.
Presented with Wassim Chegham at Best of Web Paris 2018.

Guillaume Laforge

June 07, 2018
Tweet

More Decks by Guillaume Laforge

Other Decks in Technology

Transcript

  1. - Git - Shell - Node + NPM - Code

    Editor - Firebase + Github Accounts Requirements
  2. 87% of US B2C marketers believe chatbots and virtual assistants

    will play a significant marketing role —eMarketer 33% of US internet users will use voice assistants —eMarketer >85% of customer interactions will be managed without a human —Gartner Conversational experiences are the future By 2019 By 2020 By 2021
  3. Answer FAQs Troubleshoot issues Resolve payment issue Confirm payment Show

    order status Change user settings Retrieve reservations or bookings Provide login or password help Help customers instantly Manage refunds/exchanges
  4. Make transactions fast and seamless Book flights Schedule rides Purchase

    products Buy tickets Order food Send order updates Purchase subscriptions Book appointments or reservations Recommend other goods or services
  5. Delight customers in fun and useful ways Start car Learn

    new things Turn on lights Lock house Play games Get directions Recommend recipes Suggest workout Play music, video, news, podcast
  6. 40 pre-built agents and small talk features Go-to-market in hours

    or days Start training with only a few examples Build faster
  7. High accuracy intent recognition, context awareness, slot filling Advanced natural

    language understanding and machine learning Training and analytics across platforms Engage efficiently
  8. Life of a conversation “Ok Google, talk to Air France”

    Invoke “Air France” action “Hi! Would you like to book a trip?” Speech to Text “Yes, I’d like to fly from Paris to SFO tomorrow” Text to Speech “Sure, here’s Air France” Speech to Text, NLP, Knowledge Graph, ML Ranking, User Profile Text to Speech “I found one flight, AF84, tomorrow at 10:10am”
  9. Confidential & Proprietary Chatbot concepts INTENT → “search-flight” ENTITY →

    “Paris”, “SFO”, “tomorrow” Flights from Paris to SFO tomorrow? I found one flight: AF84. Want to know more?
  10. Confidential & Proprietary Chatbot concepts At what time does it

    leave? A natural conversation that learns from past exchanges CONTEXT → remember the details of the conversation Air France flight 84 departs at 10:10am FULFILLMENT → Call some business logic on a remote URL