Upload
stuart-wiener
View
137
Download
0
Embed Size (px)
Citation preview
Prototyping Chat Bots
DESIGNING THE DETAILS THAT MATTER
Stuart Wiener November 2016
HI!
DESIGNING THE DETAILS THAT MATTER
It doesn’t occur to most people that everything is designed.
~ Bill Moggridge ~
Make the right thing
Make the thing right
DON’T MAKE A PONY EXPRESS
Make the right thing
Make the thing right
• Watch and learn from end users as often as possible, for inspiration and validation.
• Build in small increments, testing the effects of every detail.
USE LEAN STARTUP TO DISCOVER SOLUTIONS THAT CREATE REAL IMPACT
BUILD
MEASURELEARN
Humans are complicated…
Confirmation Bias
We tend to believe information that we’ve heard before while ignoring
evidence that refutes our point of view.
Availability Bias
We tend to only consider evidence that
is immediately apparent to us.
Ikea Effect
We tend to place disproportionate value on the things that we make,
because, well, we made them!
Loss Aversion
We are more sensitive to potential
losses than we are to potential gains (usually by a factor of ~2)
Only when we concede that we cannot depend on our ability to predict the future are we open to a process that discovers it.
~ Duncan Watts ~
+ = ?
Nascent technology People
We expect cultural preconceptions to apply
😂
Might not mean what you think it does…
We tend to reflect other peoples emotions back to them
HI!
The context frames the interaction
Our interactions change as we develop a shared history
Our interactions change as we learn more about someone
WOAH! Chat bots have got a little way to go…
Disruption on the first day always looks like a toy…
~ Steve Blank ~
Most people think first of what they want to express or make, then find the audience for their idea.
You must work the opposite angle, thinking first of the public. You need to keep your focus on their changing needs…
Beginning with their demand, you create the appropriate supply.
~ Curtis Jackson (50 Cent) ~
Customer discovery
Discover customer needs that
could be addressed with a chat
bot.
Frame the problem
Create a problem statement that
frames the problem in the right
way—to help make sure that you solve the root problem, not a
symptom of it.
NOW IT’S YOUR TURN…
Prototype a bot
Quickly test your ideas to validate
your understand of customer
needs and behaviour.
1 2 3
Remember that our higher order objective is to validate our ideas the fastest, cheapest way possible. Actually building and
launching a product idea is generally the slowest, most expensive way to validate the idea.
~ Marty Cagan ~
Pay attention to what users do, not what they say.
~ Jacob Nielsen ~
• What is the job that a customer will hire your product to do?
• Tasks to complete
• Problems to solve
• Needs to satisfy
• What are the pains involved?
• What gains could be made even better?
UNDERSTAND THE JOBS-TO-BE-DONE
EXISTING BEHAVIOUR
NEW BEHAVIOUR
Attraction of new solution
Attraction of old solution . Push from new solution
Push from old solution
EXISTING BEHAVIOUR
NEW BEHAVIOUR
Motivation Beliefs, perceptions
Carrot & stick
Social norms
Ability Knowledge
Skill (usability)
Habit
Opportunity Trigger Timing
Context
The forces behind those…
• What did people do?
• When did they do those things?
• What was difficult?
• What was easy?
• Even better if they can show you, rather than tell you…
START WITH STORIES
What humans are best at doing is interpreting all new information so that their prior conclusions are left intact.
~ Warren Buffett ~
• Ask about specific events.What do you normally do? What did you do yesterday?
• Ask open questions—don’t preempt the answer.Did you like that? What do you think about that?
• Ask about emotional impactHow does that make you feel?
• Drill down to root causesWhy? Why? Why?
GETTING WORTHWHILE INSIGHTS FROM INTERVIEWS
Find a friend
Interview them about one of their hobbies for 5 minutes each
You’re looking for pains, gains & jobs-to-be-done
10 mins
How did it go? Did you learn anything interesting?
Did you get pains, gains & jobs-to-be-done?
Finding the problem is the hardest part.
~ Kevin Systrom ~
Work alone Write a few problem statements based on what you learnt
5 mins
How might we help _______________ to
__________________ given that __________________.your friend
job-to-be-donepain / gain
How did it go? Can you imagine a bot that could help?
A prototype is a quick test of a possible design solution. When prototyping a chat bot it’s particularly easy. All you need is…
• A friend
You will “be the bot”.
EXPLORATORY PROTOTYPING
Push: without being prompted you’re going to message your friend. You might need to set the scene for them by giving them a little context…
Pull: your friend is going to message you first, after which you, the bot, will reply.
WHAT SORT OF BOT ARE YOU?
Spend 2 minutes thinking about how the conversation might go (what information are you going to need readily available?)
Spend 3 minutes having a WhatsApp conversation—try not to talk verbally to each other.
Try to stress test your friend—use realistic but extreme scenarios if possible.
10 mins
How did it go? What did you learn?
What was different to what you expected or planned?
KEEP GOING…BUILD
MEASURELEARN