Welcome to ServerlessToronto.org
1
Sep 14, 2021: “Get started with Dialogflow & Contact
Center AI on Google Cloud” with Lee Boonstra.
Introduce yourself if you “arrive” before we
start at 12:05pm ☺
Serverless Evolution (since FaaS started)
2
Serverless is New Agile & Mindset
#1 We started as Back-
end FaaS (Serverless)
Developers who enjoyed
“gluing” other people’s
APIs and Managed
Services)
#3 We're obsessed by
creating business value
(meaningful MVPs,
Products), focusing on
Outcomes – NOT
Outputs, and we mesh
well with Product
Managers
#2 We build bridges
between Serverless
Community (“Dev leg”),
and Front-end, Voice-First
& UX folks (“UX leg”)
#4 Achieve agility NOT
by “sprinting” faster
(like in Scrum), but
working smarter (by
using bigger building
blocks and less Ops)
Disconnect between IT & Business needs
3
How to help companies accelerate?
Technology is not the point => We are here to create Value
Adopting Serverless Mindset allowed us to shift the focus from “pimping
up our cars” (infrastructure/code), towards “driving” (the business) forward.
≠
Let’s bridge the Businesses & IT Gap by:
4
1. bringing more Business-focused
topics (like one today) to educate,
2. offering free Second Opinions on
Application/Data Architecture
modernization (to Businesses),
3. offering for-fee Consulting service
(regardless of how short they are),
4. connecting Cloud enthusiasts from
the Community with Employers
Fill the Survey help us serve you better,
plus to win Lee’s NEW book:
https://forms.gle/oH2ZTnSgMTH41xsg7
Upcoming ServerlessToronto.org Meetups
5
1) Dr. Maloy – Empowering
Developers to be Healthcare
Heroes
2) Intro to Vertex AI, unified MLOps
platform for Data Scientists & ML
Engineers – Jarek Kazmierczak &
Brian Kang from Google
3) An Evening with Felipe Hoffa –
Big Data guru from Snowflake
4) YOUR “This is my Architecture”
style presentations are welcome!
Regardless how big or small
your learning & sharing will be ☺
Please rate us on Meetup, Tell your peers
We’re here to Help YOU help others
6
Feature Presentation
Lee Boonstra | @ladysign
Get started with Dialogflow &
Contact Center AI on Google Cloud
Lee Boonstra
Developer Relations Engineer, Google
Lee Boonstra | @ladysign
Lee Boonstra
Developer Relations Engineer, Conversational AI
@Google
Author of the book: The Definitive Guide to
Conversational AI with Dialogflow & Google Cloud
(Apress)
https://www.leeboonstra.dev
@ladysign
Proprietary + Confidential
Agenda
1
2
3
4
Intro Conversational AI Google
Dialogflow Different Versions
Contact Center AI
My Omnichannel Project
Code Repos
Q&A
5
6
Lee Boonstra | @ladysign
Chatbots exists since the 90’s…
So why are they popular now?
FIRST CHATBOT, 1994
SHIFT FROM A MOBILE FIRST
TO AN AI FIRST STRATEGY
Lee Boonstra | @ladysign
Chatbots.
It's all about Machine Learning.
● Natural Language Understanding
● Intent Matching
● Speech to Text
● Text to Speech (Wavenet)
Lee Boonstra | @ladysign
Main Categories of Conversational AI
IoT
Search Customer
Care
Assistance
Lee Boonstra | @ladysign
Conversational AI as part of Google Cloud
Dialogflow ES
Dialogflow CX
Speech to Text Text to Speech
Insights
CCAI Agent Assist
Developer APIs
SAAS
Partner GTM
offering
Lee Boonstra | @ladysign
Dialogflow usage
● 2016 - 2020 - people started to
experience with conversational
AI
● Our conversational UI
development suite: Dialogflow,
has a user base of more than
1.7M.
Development Suit for
building Conversational UIs
Lee Boonstra | @ladysign
What we see at Google
In the beginning (2016), most conversational AIs
● Simple chatbots / voice bots - 1 or 2 turn taking turns
● "Example Google Assistant bot;
○ 'hey Google, what's my balance?'
Building conversational AIs is an ongoing process, and thus the
conversational AIs become more complicated over the years.
Lee Boonstra | @ladysign
Chatbots are expected to cut
business costs by more than
$8 billion per year by 2022
Juniper Research
Lee Boonstra | @ladysign
Lee Boonstra | @ladysign
What we see at Google
● The big $$$ savers are typically
complex use-cases
● Replacement or add-ons for Customer
Care / contact centers
○ Which is very expensive for
businesses
○ With conversational AI we can:
■ Reducing Calltime,
■ Reducing hold time
■ Offer 24/7 availability
Lee Boonstra | @ladysign
The next generation of conversational
UI's are more complex.
● Large complex flows / giant implementations
○ hundreds or thousands of intents
● More than 3 turn taking turns (branching of conversations)
● Parts of a dialogue are repeatable
● Understanding the intent & context from long utterances (in
contact centers)
● Working with teams on large implementations
● Maintaining an always growing conversation
Lee Boonstra | @ladysign
Introducing
Development suite for building large &
complex Conversational UIs.
BETA
An additional conversational
product which co-exist next to
the other Dialogflow tooling…
● Build complex enterprise
conversations
● Easier to maintain
DEMO
Lee Boonstra | @ladysign
60%
want easier access to
self-service solutions
prefer personal interactions, not
necessarily with humans
Note: US statistics millennials; Source: Zendesk, Statista, Strategy Analytics, HBR, Ovum, USA market, CISCO VNI
expect to receive
real-time assistance
64% 75%
Customers expect great, flexible, & personal experiences
Lee Boonstra | @ladysign
Often the reality looks like this
● Long waiting / hold times
● Unlimited Call transfers
● IVR difficult to navigate
● Availability
● Inadequate information
● Agents have to answer
same types of questions
over and over again.
https://www.youtube.com/watch?
v=72tF7ggdQyU
Lee Boonstra | @ladysign
Introducing Google Cloud Contact Center AI
Out of the box solution for telephony
contact centers to use AI. No machine
learning expertise required.
It's using popular Google Cloud components...
Lee Boonstra | @ladysign
Use machine intelligence, to make the
experience of calling a service number
better; better for the customer,
and better for the agent
Lee Boonstra | @ladysign
Bots in a telephone contact center are actually
much more complicated...
Turn on the lights
Short utterance. -> Match to 1 intent. Long utterance. Many possible
intent matches.
Lee Boonstra | @ladysign
...and speech quality is critical
● Phone Quality: 8kz mono
● Background noises
● Speaker Diarization (multiple speakers / interrupting each other)
● Requires special IVR features
Conversational
core
Insights
V
i
rt
u
a
l
A
g
e
n
t
A
g
e
n
t
A
s
s
i
s
t
We enable rich
interactions with
virtual agents.
Allows customers to have multi-topic and
multi-channel conversations. Bot
interactions provide more natural
experiences. Multi-lingual support.
We make agents
superheroes in
the eyes of
customers.
Agents can quickly respond to customers with the suggested
responses rather than needing to determine the right
response themselves. Agents are suggested knowledge base
content to solve a customer’s issue so they don’t have to ask a
customer to wait while they search.
We offer the
visibility and
actionable ways to
improve contact
center performance.
Analysts no longer need to manually review and categorize
conversations. Accurate and scaled coverage of call driver
assignment allows easy data segmentation and helps drive
deeper analysis of conversations. Data from Virtual Agents
and Agent Assist feed Insights, providing continuous
evaluation and cyclical feedback.
Proprietary + Confidential
Cloud
STT
Session Manager
State Management
Agent Assist
Articles
Cloud TTS
(Wavenet)
Cloud STT
Full Session Transcripts
Dialogflow
Virtual Agent
Agent Assist
NLU + Model
Knowledge
service/KB
Agent Assist
Feedback
Partner Virtual Agent Path Human Agent Path
Agent
Desktop
Telephony
Partner
Customer
Call
General Reference Architecture
(Voice)
Lee Boonstra | @ladysign
What's possible with
Contact Center AI
in combination with Genesys IVR
system
https://www.youtube.com/watch?v=
hWmPLOJPYSc
Lee Boonstra | @ladysign
A preview of
CCAI Insights
Lee Boonstra | @ladysign
Conversational AI Summary
● Developer APIs, such as Speech-to-Text, Text-to-Speech, Natural Language Understanding.
● Conversational AI suites / SAAS:
○ Dialogflow Essentials
○ Dialogflow CX (advanced conversational AIs)
○ Insights AI (suite for detecting the call drivers of a conversation)
● Contact Center AI & Agent Assist
○ GTM offering, where Google delivers the above components + specific Contact Center
APIs so telephony partners can integrate this in their existing architectures.
Lee Boonstra | @ladysign
Lee Boonstra | @ladysign
Future of Customer Care
● Proactive
● Ubiquitous
● Omnichannel
"Make your call center not a cost center but a revenue center."
Lee Boonstra | @ladysign
An example of
a conversational
omnichannel
experience
Lee Boonstra | @ladysign
The technology is already there
RCE 360 Architecture
Twilio
Telephony
Backend Container
Dialogflow
ES
Stackdriver
Cloud
Pubsub
Twilio
SMS
Mobile
App
Google
Assistant
Dialogflow
CX
Nginx Container
Website
Container
Server
Actions on Google
Middleware
Flutter
Middleware
Twilio
Middleware
Website
Middleware
NLU
Cloud
Functions
AdLingo
Dialogflow
ES
External
Load
Balancer
Lee Boonstra | @ladysign
Code Repos
● https://github.com/savelee/kube-django-ng
○ http://www.futurebank.nl
● https://github.com/dialogflow/selfservicekiosk-audio-streaming
○ https://selfservicedesk.appspot.com/
Lee Boonstra | @ladysign
Thank you!
Twitter: ladysign
Amazon: Lee Boonstra
https://www.leeboonstra.dev/chatbots/dialogflow-the-book/
Proprietary + Confidential
Join www.ServerlessToronto.org
Home of “Less IT Mess”

Get started with Dialogflow & Contact Center AI on Google Cloud

  • 1.
    Welcome to ServerlessToronto.org 1 Sep14, 2021: “Get started with Dialogflow & Contact Center AI on Google Cloud” with Lee Boonstra. Introduce yourself if you “arrive” before we start at 12:05pm ☺
  • 2.
    Serverless Evolution (sinceFaaS started) 2 Serverless is New Agile & Mindset #1 We started as Back- end FaaS (Serverless) Developers who enjoyed “gluing” other people’s APIs and Managed Services) #3 We're obsessed by creating business value (meaningful MVPs, Products), focusing on Outcomes – NOT Outputs, and we mesh well with Product Managers #2 We build bridges between Serverless Community (“Dev leg”), and Front-end, Voice-First & UX folks (“UX leg”) #4 Achieve agility NOT by “sprinting” faster (like in Scrum), but working smarter (by using bigger building blocks and less Ops)
  • 3.
    Disconnect between IT& Business needs 3 How to help companies accelerate? Technology is not the point => We are here to create Value Adopting Serverless Mindset allowed us to shift the focus from “pimping up our cars” (infrastructure/code), towards “driving” (the business) forward. ≠
  • 4.
    Let’s bridge theBusinesses & IT Gap by: 4 1. bringing more Business-focused topics (like one today) to educate, 2. offering free Second Opinions on Application/Data Architecture modernization (to Businesses), 3. offering for-fee Consulting service (regardless of how short they are), 4. connecting Cloud enthusiasts from the Community with Employers Fill the Survey help us serve you better, plus to win Lee’s NEW book: https://forms.gle/oH2ZTnSgMTH41xsg7
  • 5.
    Upcoming ServerlessToronto.org Meetups 5 1)Dr. Maloy – Empowering Developers to be Healthcare Heroes 2) Intro to Vertex AI, unified MLOps platform for Data Scientists & ML Engineers – Jarek Kazmierczak & Brian Kang from Google 3) An Evening with Felipe Hoffa – Big Data guru from Snowflake 4) YOUR “This is my Architecture” style presentations are welcome! Regardless how big or small your learning & sharing will be ☺ Please rate us on Meetup, Tell your peers We’re here to Help YOU help others
  • 6.
  • 7.
    Lee Boonstra |@ladysign Get started with Dialogflow & Contact Center AI on Google Cloud Lee Boonstra Developer Relations Engineer, Google
  • 8.
    Lee Boonstra |@ladysign Lee Boonstra Developer Relations Engineer, Conversational AI @Google Author of the book: The Definitive Guide to Conversational AI with Dialogflow & Google Cloud (Apress) https://www.leeboonstra.dev @ladysign
  • 9.
    Proprietary + Confidential Agenda 1 2 3 4 IntroConversational AI Google Dialogflow Different Versions Contact Center AI My Omnichannel Project Code Repos Q&A 5 6
  • 10.
    Lee Boonstra |@ladysign Chatbots exists since the 90’s… So why are they popular now?
  • 11.
  • 12.
    SHIFT FROM AMOBILE FIRST TO AN AI FIRST STRATEGY
  • 13.
    Lee Boonstra |@ladysign Chatbots. It's all about Machine Learning. ● Natural Language Understanding ● Intent Matching ● Speech to Text ● Text to Speech (Wavenet)
  • 14.
    Lee Boonstra |@ladysign Main Categories of Conversational AI IoT Search Customer Care Assistance
  • 15.
    Lee Boonstra |@ladysign Conversational AI as part of Google Cloud Dialogflow ES Dialogflow CX Speech to Text Text to Speech Insights CCAI Agent Assist Developer APIs SAAS Partner GTM offering
  • 16.
    Lee Boonstra |@ladysign Dialogflow usage ● 2016 - 2020 - people started to experience with conversational AI ● Our conversational UI development suite: Dialogflow, has a user base of more than 1.7M. Development Suit for building Conversational UIs
  • 17.
    Lee Boonstra |@ladysign What we see at Google In the beginning (2016), most conversational AIs ● Simple chatbots / voice bots - 1 or 2 turn taking turns ● "Example Google Assistant bot; ○ 'hey Google, what's my balance?' Building conversational AIs is an ongoing process, and thus the conversational AIs become more complicated over the years.
  • 18.
    Lee Boonstra |@ladysign Chatbots are expected to cut business costs by more than $8 billion per year by 2022 Juniper Research Lee Boonstra | @ladysign
  • 19.
    Lee Boonstra |@ladysign What we see at Google ● The big $$$ savers are typically complex use-cases ● Replacement or add-ons for Customer Care / contact centers ○ Which is very expensive for businesses ○ With conversational AI we can: ■ Reducing Calltime, ■ Reducing hold time ■ Offer 24/7 availability
  • 20.
    Lee Boonstra |@ladysign The next generation of conversational UI's are more complex. ● Large complex flows / giant implementations ○ hundreds or thousands of intents ● More than 3 turn taking turns (branching of conversations) ● Parts of a dialogue are repeatable ● Understanding the intent & context from long utterances (in contact centers) ● Working with teams on large implementations ● Maintaining an always growing conversation
  • 21.
    Lee Boonstra |@ladysign Introducing Development suite for building large & complex Conversational UIs. BETA An additional conversational product which co-exist next to the other Dialogflow tooling… ● Build complex enterprise conversations ● Easier to maintain
  • 22.
  • 23.
    Lee Boonstra |@ladysign 60% want easier access to self-service solutions prefer personal interactions, not necessarily with humans Note: US statistics millennials; Source: Zendesk, Statista, Strategy Analytics, HBR, Ovum, USA market, CISCO VNI expect to receive real-time assistance 64% 75% Customers expect great, flexible, & personal experiences
  • 24.
    Lee Boonstra |@ladysign Often the reality looks like this ● Long waiting / hold times ● Unlimited Call transfers ● IVR difficult to navigate ● Availability ● Inadequate information ● Agents have to answer same types of questions over and over again. https://www.youtube.com/watch? v=72tF7ggdQyU
  • 25.
    Lee Boonstra |@ladysign Introducing Google Cloud Contact Center AI Out of the box solution for telephony contact centers to use AI. No machine learning expertise required. It's using popular Google Cloud components...
  • 26.
    Lee Boonstra |@ladysign Use machine intelligence, to make the experience of calling a service number better; better for the customer, and better for the agent
  • 27.
    Lee Boonstra |@ladysign Bots in a telephone contact center are actually much more complicated... Turn on the lights Short utterance. -> Match to 1 intent. Long utterance. Many possible intent matches.
  • 28.
    Lee Boonstra |@ladysign ...and speech quality is critical ● Phone Quality: 8kz mono ● Background noises ● Speaker Diarization (multiple speakers / interrupting each other) ● Requires special IVR features
  • 29.
    Conversational core Insights V i rt u a l A g e n t A g e n t A s s i s t We enable rich interactionswith virtual agents. Allows customers to have multi-topic and multi-channel conversations. Bot interactions provide more natural experiences. Multi-lingual support. We make agents superheroes in the eyes of customers. Agents can quickly respond to customers with the suggested responses rather than needing to determine the right response themselves. Agents are suggested knowledge base content to solve a customer’s issue so they don’t have to ask a customer to wait while they search. We offer the visibility and actionable ways to improve contact center performance. Analysts no longer need to manually review and categorize conversations. Accurate and scaled coverage of call driver assignment allows easy data segmentation and helps drive deeper analysis of conversations. Data from Virtual Agents and Agent Assist feed Insights, providing continuous evaluation and cyclical feedback.
  • 30.
    Proprietary + Confidential Cloud STT SessionManager State Management Agent Assist Articles Cloud TTS (Wavenet) Cloud STT Full Session Transcripts Dialogflow Virtual Agent Agent Assist NLU + Model Knowledge service/KB Agent Assist Feedback Partner Virtual Agent Path Human Agent Path Agent Desktop Telephony Partner Customer Call General Reference Architecture (Voice)
  • 31.
    Lee Boonstra |@ladysign What's possible with Contact Center AI in combination with Genesys IVR system https://www.youtube.com/watch?v= hWmPLOJPYSc
  • 32.
    Lee Boonstra |@ladysign A preview of CCAI Insights
  • 33.
    Lee Boonstra |@ladysign Conversational AI Summary ● Developer APIs, such as Speech-to-Text, Text-to-Speech, Natural Language Understanding. ● Conversational AI suites / SAAS: ○ Dialogflow Essentials ○ Dialogflow CX (advanced conversational AIs) ○ Insights AI (suite for detecting the call drivers of a conversation) ● Contact Center AI & Agent Assist ○ GTM offering, where Google delivers the above components + specific Contact Center APIs so telephony partners can integrate this in their existing architectures.
  • 34.
    Lee Boonstra |@ladysign
  • 35.
    Lee Boonstra |@ladysign Future of Customer Care ● Proactive ● Ubiquitous ● Omnichannel "Make your call center not a cost center but a revenue center."
  • 36.
    Lee Boonstra |@ladysign An example of a conversational omnichannel experience
  • 37.
    Lee Boonstra |@ladysign The technology is already there RCE 360 Architecture Twilio Telephony Backend Container Dialogflow ES Stackdriver Cloud Pubsub Twilio SMS Mobile App Google Assistant Dialogflow CX Nginx Container Website Container Server Actions on Google Middleware Flutter Middleware Twilio Middleware Website Middleware NLU Cloud Functions AdLingo Dialogflow ES External Load Balancer
  • 38.
    Lee Boonstra |@ladysign Code Repos ● https://github.com/savelee/kube-django-ng ○ http://www.futurebank.nl ● https://github.com/dialogflow/selfservicekiosk-audio-streaming ○ https://selfservicedesk.appspot.com/
  • 39.
    Lee Boonstra |@ladysign Thank you! Twitter: ladysign Amazon: Lee Boonstra https://www.leeboonstra.dev/chatbots/dialogflow-the-book/
  • 40.
  • 41.