1. Session code

2. Bots as the Next UX: Expanding Your Apps with Conversation
Machine Learning & Data Science Conference
9/4/2018 7:49 PM
Bots as the Next UX: Expanding Your Apps with Conversation
Matt Winkler | GPM, Data Group
© 2015 Microsoft Corporation. All rights reserved. MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION IN THIS PRESENTATION.

3. Agenda
What
Why
Why Not
Why Now
How

4. What’s a Bot?
Bots are services that people interact with through conversation and messaging

5. Bots are not… AI Machine Learning Self aware, that we know of
A replacement for every UX we use today

6. Why Bots? Meet customers where they are at
Communicate with customers the way they do
Interact through the apps your customers use today
More personal than an app
Develop deeper connections

7. Why now? People spend more time messaging
Increase in capability/focus in messaging platforms
Rise in accessible language processing technology
We have reached “peak app”

8. Why not bots? My users want to type everything!
I can replace my other channels for customer interaction!
If I build a bot, profit!

9. So, you want to build a bot?
Design the interaction
Writing the code
Host the code
Connect to it
Learn
Repeat

10. Designing the Interaction
Basic request response
“Info” scenarios
“Helper” scenarios
Basic Hierarchy
“Select” scenarios
Minimal state stored
Nested waterfalls
Configuration/multi-step
Requires state + reset

11. Designing the Interaction
Arbitrary state machine
“Free form” scenarios
Requires navigation for the user

12. Show Me Some Bots
Memecat
Azure bot
Hipmunk

13. Authoring the code C# and Node.js Dialogs to model a conversation
Dialogs are reusable
Conversations are scalable to multiple machines
Dialog state is persisted in Bot Connector (per-user, per-session, per- user+session)
Types of Dialogs:
Built-in prompts
Yes/No, String, Number, Choices
Natural language understanding (LUIS)
Form slot filling (branching, disambiguation, multi-turn)
Profile (e.g., home address)

14. Coding

15. Hosting Your Bot All you need is compute! May we recommend:
Azure App Services
Azure Virtual Machines
Azure Functions
You may also use your own hosting
(needs internet routable URL)

16. Hosting

17. Connect It Up Just the web control Skype, it just works!
Facebook, Slack, and more
Going direct for full control

18. Bot Connector Basics

20. Bot Connectors

21. Actively Learning with LUIS
LUIS – Language Understanding Intelligent Service
Enables natural language understanding
Augments more complex dialog flows
Balance between guiding and open ended
Consider the global exception handler

22. Language understanding intelligent service
9/4/2018 7:49 PM
Language understanding intelligent service
Reduce labeling effort with interactive featuring Use visualizations to gauge performance and improvements Leverage speech recognition with seamless integration Deploy using just a few examples with active learning
Define
concepts
Provide
examples
Active learning
Deploy

23. Language understanding models
9/4/2018 7:49 PM
Language understanding models {
“entities”: [
“entity”: “flight_delays”,
“type”: “Topic” } ],
“intents”: [
“intent”: “FindNews”,
“score”: },
“intent”: “None”,
“score”:
“intent”: “ReadNews”,
“score”:
“intent”: “ShareNews”,
“score”: E-06 ]
“News about flight delays”

25. LUIS

26. LUIS

27. What’s Next? Write your first bot!
Consider ways conversation augments your apps
Register your bot in the bot directory
Join the conversation!

28. Learning More Bot Framework – http://dev.botframework.com
LUIS –
Sources –
AzureBot –

29. 9/4/2018 7:49 PM
© 2015 Microsoft Corporation. All rights reserved. MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION IN THIS PRESENTATION.