3. Conversation UIs & Chatbots for your Digital Workplace
• Introduction (who am I)
• What are conversational UIs and what’s the difference with AI and AR
• Why should you want to start with it
• How getting started
4. Pulse Check
• Who already has a chatbot/ virtual assistant in your organisation?
• Aimed at clients/ colleagues?
• Are you actively involved?
• Who is considering a chatbot/ working on it?
• What are your expectations;
What would you like to hear/ know more about today?
5. Marion Mulder
I help organisations leverage digital
technology, especially Chatbots,
Virtual Assistant, AI & AR,
for optimal Customer Service*
With people from your organization
and/or professionals from my network
Bringing an outside-in perspective
www.muldimedia.com
marion@muldimedia.com
+31642111245
*) Customer Service = how we serve our customers,
not just the helpdesk!
“Making digital technology work for you”
How can I help?
13. To complex anticipatory advisory bots
Welcome home Marion,
I have turned on the heating,
dinner is almost ready
and I have recorded Nieuwsuur for you as
there was an item on AI on it.
Would you like me to play that for you now
while you are waiting for dinner?
18:30
14. Types of bots
14
Informational
Transactional
Advisory
Informational chatbots are the
simplest type.
They usually involve providing
general information such as
FAQs, news stories and push
notifications
Transactional chatbots allow
users to complete transactions
and interact (such as booking a
hotel)
Typically they require a user to
be authenticated into their user
account
Self-learning chatbots are the
next evolution in chatbots.
They are able to learn based on
customer interactions to
determine the appropriate next
steps.
KnowledgeComplexity
AI Maturity
15. Why
• The market (customers and users will expect it from you)
• Making life (at work) easier
16. Why should you care about Conversational UIs
Customer Central Day after tomorrow ready
Business Process Optimisation
Make live at work easier
• Faster than real time customer service
• Hyper personalisation
• Convenience
• DEX; Digital Workplace – Time & Space
• Employee Self Service -> Employee Self Service
• Convenience
23. Not just design the happy flow conversation….
Small talk
The actual dialog
flow
Don’t know...
Getting back on track
Hand-off to Human
24. Not just design the happy flow conversation….
24 https://developers.google.com/actions/downloads/design-principles-quick-reference.pdf
25. Which event /action/ trigger
and
Which information need
Retrieve
where
Store
where
‘Present’
where
(Re-)mapping customer journey, with interface, content and systems
Intent
26. “I need to go to New York next week”
(ps: why did [Siri] not see this in my calendar yet and anticipated my needs?)
27. Intelligent Content - Structured Authoring and Micro Content
Source: information Energy 2018 Conference
28. • Amelia was manually trained on two basic flows in
PowerPoint and 77 knowledge articles.
• Based on this, User Stories were defined which
resulted into
• 14 Dialogue plans
• 32 Grammars (Goals & slots)
• 33 BPMN process flows
• 5 general
• 28 MobileIron specific
28
“I want company email on my private mobile phone" (MobileIron)
29. Content Model – Content Information Types
Reference
• Describes things the reader needs to
KNOW
Concept
• Explains things the reader needs to
UNDERSTAND
Task
• Instructs the reader HOW TO DO
things
Principle
• Advices the reader about what they
need TO DO or NOT DO and WHEN
Process
• Demonstrates to the reader how
things WORK
Information needs to be typed according to the
intended reader response to that content
32. “paper prototype”: fake it
• Have your script and default (designed) answers ready
• Put a actual person on the receiving end but tell the
test user that it’s a bot
• Play it out
• Record the conversation
• If successful use the recorded data to train your
actual bot
• If not successful learn where you need to
improve (had to go off script)
33. Suggest speaker conversation flow (or input form)
What is the speaker’s name?
She is called [firstname] [lastname]
[expertise], [expertise], [expertise],
[expertise]
What topics can we #AskHer about as
a speaker or expert?
Thanks you
You have provided the
following information
Name: [firstname [lastname]
Experise: [topic], [topic]
Organsation [job] [organsation]
Identifies as: [identity, identiy]
More infro: [url]
Photo [show photo]
Is this correct?
She is [job title] at
[organization]
What is the speaker’s organisation?
[www.linkedin.com/name
ofperson/]
Is there a website or linkedin profile
we could refer to for more info?
If so please share URL
[woman]
Do you know how she identifies (e.g.
black, moroccan, lesbian, etc)
[upload] or [photo url]
Can you provide a photo (upload or
share photo source url)
skip
skip
skip
skipupload Add link
submitChange
[name] has been added to
our list. We will contact
her.
Can we use you as a
reference?
No thanksyes
Would you like to add
someone else?
No thanksyes
Find expert/
speaker
Suggest
speaker
34. What if someone adds all or most of the content in 1 response?
What is the speaker’s name?
I would like to suggest [firstname]
[lastname] she works at
[organization] as a [jobtitle] and you
should ask her about [topic], [topic]
you can find more info about here
at [link]
Suggest
speaker
35. New employee onboarding (Digital workplace only)
I have a new
employee
Is it a permanent
employee or
temporary?
PermanentTemporary
[ab12cd]
Temps do not get
laptop or phone.
They use remote
access on their own
computer.
Request remote
access for new
employee?
yesNo
Do you have [corpID]
yet?
No
Shall I order laptop
for [ab12cd]?
yesNo
Shall I order the
standard phone with
company mobile
number [ab12cd]?
Or will they bring
their own device?
Nr onlyNo yes
37. Implement & Scale
• Start with simple processes or FAQs, preferably with high volume or
high impact
• Think about how the new interface will fit your (future) landscape
• Think about how to keep your content up to par and how to fit it in
your content landscape
38. How to build a Virtual Assistant
Define Design Flow Design
Conversation
User Testing!! Implement
& Scale
To achieve good ‘Intent’
classification accuracy, it’s
important to provide your
engine with enough data.
The greater is the number of
natural language examples in
the ‘User says’ section of
intents, the better is the
classification accuracy.
Sourcing “User says”
• Current human chat logs
• Emailed questions
• Other sources
• Wizard of Oz/ Mechanical
Turk
• Define its purpose (value
proposition)
• Pick the right cases
• Gather input on the cases
(“epics’)
• Define intents
• Define measures of
success
• What backend systems or
3rd party services need to
be connected
• Define chatbot needs a
personality
Tools
• Design Sprint
• Chatbot Design Canvas
• What steps/task will be
needed in the conversation;
both high level flow and
detailed flows
• What is fastest route from
A-Z (is not your web flow)
• Not just happy flow; people
WILL deviate
• Provide escapes (e.g.
handover to human agent)
• Impact of process redesign
on organisation
Tools
• Customer Journeys
• Existing process flows
• Input from SMEs (e.g.
helpdesk)
• Don’t pretend to be human
• Basic greetings and goodbyes
• Tone of voice based on chatbot
persona
• Conversational dialogs
• Conversation repairs
• Balance between text and rich
media (buttons, images, videos,
links, smart syntax hints)
• Balance between re-writing
everything vs referring to or
presenting other sources
(Intelligent Content Model)
Tools
• Google Conversation Design
checklist
Example platforms
• DigitalCX
• Dialogflow (google)
• IBM Watson
• IP Soft Amelia
• etc
Data privacy (CIA, BIA)
Machine learning capabilities
Frontend integrations
Back-end, API’s and
connecting systems/
databases
Bot-to bot readiness
CMS alignment
Build a prototype which
allows you to user test the
conversation.
You can build different
prototypes in different ways
for different purposes
If you have a really good
training set your 1st user test
can be successful
But brace yourself….
User will ask questions in
ways you couldn’t have
imagined!
Use the user test data to
improve your conversation
flow and training data
This will be an ongoing
process
Prototype (create)
Training (Data)