Talk from "We Are Developers World Congress" Session Info: Introduction into the Open Source Framework "E.D.D.I", that has been developed for creating and maintaining multiple Chatbot-Products in a Cooperate Environment. This talk will cover the architecture, how it can be used and how it has been used, based on an example with the Norwegian Company "differ.chat". Speaker Bio: Gregor Jarisch Chatbot development since 2006 (industries such as e-commerce, first-level-support, quality management, education). Chatbot Lead @ DIFFER.CHAT Dev Lead of Enterprise-Ready Open Source Chatbot Platform "E.D.D.I.". Agile and Innovation Coaching 10+ years work experience in Software Development, in particular Web Services.

1. E.D.D.I
Open Source
Chatbot Platform

2. Hello!
I am Gregor Jarisch
E.D.D.I
Dev Lead
DIFFER
Chatbot Lead
You can find me at:
www.jarisch.net
www.labs.ai | www.differ.chat
2

3. 1.
What is E.D.D.I ?
Short intro
3

4. E.D.D.I – Open Source Chatbot Platform
◎Enterprise-Ready
◎Multiple Chatbots with Shared Knowledge
◎NLP, Behavior, API Connector, Templating
◎Extension Oriented Plugin System
4

5. 21kLines of Code
4Development Iterations from Scratch
8Years of Engineering
5

6. Our Philosophie
Flexibility /
Adaptability
Structure /
Reusability
Performance /
Efficency
6

7. 2.
Example in Action
Let’s Build a Simple
Weather Feature
7

8. “
What’s the weather?
8

9. “
What’s the weather?
9
Which city would you like to know
the weather of?

10. “
Vienna
10

11. “
Vienna
11
The current weather situation in
Vienna is cloudy at 15°C.

12. User Interaction Process
Behavior
12
Memory
Parser
Output
API
Connector

13. User Interaction Process
Behavior
13
Memory
Parser
Output
Meaning
Actions
Input
Bot Reply
API
Connector
Actions

14. User Interaction Process
Behavior
14
Memory
Parser
Output
Meaning
Actions
Input
Bot Reply
API
Connector
Actions
“What’s the weather”

15. User Interaction Process
Behavior
15
Memory
Parser
Output
Meaning
Actions
Input
Bot Reply
API
Connector
Actions
“What’s the weather”
trigger(current_weather)

16. User Interaction Process
Behavior
16
Memory
Parser
Output
Meaning
Actions
Input
Bot Reply
API
Connector
Actions
“What’s the weather”
trigger(current_weather)
fetch_weather,
current_weather

17. User Interaction Process
Behavior
17
Memory
Parser
Output
Meaning
Actions
Input
Bot Reply
API
Connector
Actions
“What’s the weather”
trigger(current_weather)
fetch_weather,
current_weather
fetch_weather

18. User Interaction Process
Behavior
18
Memory
Parser
Output
Meaning
Actions
Input
Bot Reply
API
Connector
Actions
“What’s the weather”
trigger(current_weather)
fetch_weather,
current_weather
current_weather
fetch_weather

19. User Interaction Process
Behavior
19
Memory
Parser
Output
Meaning
Actions
Input
Bot Reply
API
Connector
Actions
“What’s the weather”
trigger(current_weather)
fetch_weather,
current_weather
current_weather
fetch_weather
“The current weather situation
in Vienna is cloudy at 15°C.”

20. 3.
Bloody Code
How does it look like for real
20

21. Dictionary
21
{
"words": [
{
"word": "weather",
"exp": "trigger(current_weather)"
}
],
"phrases": [
{
"phrase": "whats the weather",
"exp": "trigger(current_weather)"
}
]
}

22. Dictionary
22
{
"words": [
{
"word": "weather",
"exp": "trigger(current_weather)"
}
],
"phrases": [
{
"phrase": "whats the weather",
"exp": "trigger(current_weather)"
}
]
}

23. Behavior Rule (1)
23
{
"behaviorGroups": [
{
"name": "Questions about weather",
"behaviorRules": [
{
"name": "Ask for City",
"actions": [
"ask_for_city"
],
"children": [
{
"type": "inputmatcher",
"values": {
"expressions": "trigger(current_weather)"
}
}
]
}
]
}
]
}

24. Behavior Rule (1)
24
{
"behaviorGroups": [
{
"name": "Questions about weather",
"behaviorRules": [
{
"name": "Ask for City",
"actions": [
"ask_for_city"
],
"children": [
{
"type": "inputmatcher",
"values": {
"expressions": "trigger(current_weather)"
}
}
]
}
]
}
]
}

25. {
"behaviorGroups": [
{
"name": "Questions about weather",
"behaviorRules": [
{
"name": "Current Weather in City",
"actions": [
“fetch_current_weather",
"current_weather_in_city"
],
"children": [
{
"type": "inputmatcher",
"values": {
"expressions": "trigger(current_weather)",
"occurrence": "lastStep"
}
}
]
}
]
}
]
}
Behavior Rule (2)
25

26. {
"behaviorGroups": [
{
"name": "Questions about weather",
"behaviorRules": [
{
"name": "Current Weather in City",
"actions": [
"fetch_current_weather“,
"current_weather_in_city"
],
"children": [
{
"type": "inputmatcher",
"values": {
"expressions": "trigger(current_weather)",
"occurrence": "lastStep"
}
}
]
}
]
}
]
}
Behavior Rule (2)
26

27. API Connector
27
{
"targetServer": "api.openweathermap.org/data/2.5/weather",
"httpCalls": [
{
"name": "currentWeather",
"saveResponse": true,
"responseObjectName": "currentWeather",
"actions": [
“fetch_current_weather"
],
"request": {
"method": "get",
"queryParams": {
"APPID": "<APP-TOKEN-ID>",
"units": "metric",
"q": "[[${memory.current.input}]]"
}
}
}
]
}

28. API Connector
28
{
"targetServer": "api.openweathermap.org/data/2.5/weather",
"httpCalls": [
{
"name": "currentWeather",
"saveResponse": true,
"responseObjectName": "currentWeather",
"actions": [
“fetch_current_weather"
],
"request": {
"method": "get",
"queryParams": {
"APPID": "<APP-TOKEN-ID>",
"units": "metric",
"q": "[[${memory.current.input}]]"
}
}
}
]
}
Templating Engine:

29. API Connector
29
{
"targetServer": "api.openweathermap.org/data/2.5/weather",
"httpCalls": [
{
"name": "currentWeather",
"saveResponse": true,
"responseObjectName": "currentWeather",
"actions": [
“fetch_current_weather"
],
"request": {
"method": "get",
"queryParams": {
"APPID": "<APP-TOKEN-ID>",
"units": "metric",
"q": "[[${memory.current.input}]]"
}
}
}
]
}
Templating Engine:

30. Output Templating
30
{
"outputSet": [
{
"action": "current_weather_in_city",
"outputs": [
{
"type": "text",
"valueAlternatives": [
"The current weather situation of [[${memory.current.input}]] is
[[${memory.current.httpCalls.currentWeather.weather[0].description}]] at
[[${memory.current.httpCalls.currentWeather.main.temp}]] °C"
]
}
]
}
]
}
Templating Engine:

31. Output Templating
31
{
"outputSet": [
{
"action": "current_weather_in_city",
"outputs": [
{
"type": "text",
"valueAlternatives": [
"The current weather situation of [[${memory.current.input}]] is
[[${memory.current.httpCalls.currentWeather.weather[0].description}]] at
[[${memory.current.httpCalls.currentWeather.main.temp}]] °C"
]
}
]
}
]
}
Templating Engine:

32. “
What’s the weather in Vienna?
32
The current weather situation in
Vienna is cloudy at 15°C.
http://bit.ly/EDDI_WEATHER_EXAMPLE

33. 4.
Enterprise
Use Case
Chatbot in Education Technology
to Trigger Engagement
33

34. 19x increased student engagement
3x increased course completion rates
By EdTech Foundry

35. Automate best
practice pedagogy
with chatbots
The Firm, Bachelor course, BI Norwegian Business School
Students joining groups with
help from Differ bot717
Students asking for study groups
in the open class community36
CASE STUDY
Kick-start relevant conversations and nudge
students to participate.

36. Nudge students
to complete their
assignments
50-95% of students fall behind on
their courses. But knowing who is
falling behind is difficult in big classes.
So Differ Bot automatically follows up
your students.
(In closed piloting.)

37. I’m not going to make my
marketing deadline. Where
can I get help?
Build data to understand
individual context..
..and trigger
personalized follow-up
Hi, John. You have a marketing
paper due in 3 days. Doesn’t
look like you submitted it yet.
I’m putting together a group
chat with students working on
this paper now. Want to join?

38. 5.
Commercial Support
Expert Knowledge Saves U-Turns
38

39. Open Source
> Core E.D.D.I
> NLP, Rules, Output,
API Connectors, …
> Full-fledged API
> Docker Image
> Optimized for Kubernetes
Open Source + Membership Subscription
Subscription
+ Dedicated Support
+ Training / Consulting
+ Advanced Configuration UI
+ Messaging Channel
Integrations
39

40. E.D.D.I
Hackathon
29. – 30. June 2018
WeXelerate
Praterstrasse 1, Vienna
www.chatbot-hackathon.com