Money Talks - A Look At Conversational Design Behind Olivia.AI
1. Money Talks
A Look At The Conversational Design Process In Fintech
Srikant Chari
2. Hi, I’m Srikant Chari, but please, call me Cheenu
• UX Researcher at BMC Software in
Austin, TX
• UT Dallas (BS’14, MS’16)
• Intersection of human behavior,
design, and technology
• Exploring chatbots in the enterprise
space
3. How many of you are familiar
with what a chatbot is?
4. What was the experience of
interacting with a chatbot like?
5. 2015 – The year I started designing
for a Chatbot,
called Olivia.
10. Take a look at these apps in the Fintech
Space.
11. Money is a behavioral problem, not a
math problem.
12. The Design Goals For Olivia
• Provide tips in a variety of spend
categories
• Provide tips in a friendly and
persuasive matter
• Provide tips that are actionable
for the user
15. Research On Grocery Shopping Spending Habits
• Conducted 5 interviews
• Secondary Research From Credible Sources
• Distill insights into a Journey Map
16.
17. Shopping Journey Insights
• Doesn’t have a fixed day for grocery shopping
• Goes grocery shopping without a plan
• Gets tempted very easily
• Has trouble deciding between items
26. Users saved 6-7% of monthly income
• Original: $450 monthly grocery for $4200 monthly income -> 10% of monthly
income.
• With Olivia’s Help: $250-300 for $4200 monthly income -> 6-7% of monthly
income
27. 90th Percentile Retention Rate
• High retention rate among fintech apps
• People come back to Olivia
29. A lot has changed in the
conversational design space
30. Dedicated tools for designing and building
chatbots
Botsociety IBM Watson Assistant Flow XO
To name a few…
31. Dedicated tools for measuring chatbot analytics
Botanalytics Chatbase Dashbot
To name a few…
32. Emergence Of Roles Dedicated To Conversational
Design
Conversation Designer Roles UX Writer Roles
33. Emergence Of Divisions Dedicated To
Conversational Design
Capital One’s Conversational AI Division
34. Chatbot Design Is Still In Its Formative Stage
• Still trying to solidify good design practices for chatbots
• Still trying to figure out ways to best conduct usability studies for chatbots
• Still trying to figure out how to design for personality
36. Follow Me On Social Media
• LinkedIn: Srikant “Cheenu” Chari
• Twitter: @minstrelman91
• Instagram: @cheenu.chari
37. References
Ariely, Dan. Predictably Irrational. New York: Harper Perenial, 2010. Paperback. • On Amir, Dan Ariely and Ziv Carmon (2008), “The Dissociation Between Monetary Assessment and Predicted Utility.”
Marketing Science. Vol. 27, No. 6: 1055- 1064.
Budiu, R. (2018, Nov 25th), “The User Experience Of Chatbots.” Nielsen Norman Group. Retrieved From: https://www.nngroup.com/articles/chatbots/
H.K. Baker, V. Ricciardi(2015), "Understanding Behavioral Aspects Of Financial Planning and Investing." Journal Of Financial Planning. Vol. 28, Issue 3, pp. 22-26.
Cialdini, Robert. Influence, The Psychology Of Persuasion. New York: Harper Business, 2006. Paperback.
Eyal, Nir. Hooked, How To Build Habit Forming Products. New York: Portfolio, 2014. Hardcover.
Frankel, R.S.(2017, July 20th). Millennials Are Spending A Ton On Basics. Are They Using Credit Cards Enough? Retrieved From: https://www.bankrate.com/credit-cards/spending-index-0717/
Hall, Erika. Conversational Design. New York: A Book Apart, 2018. Paperback.
Kahneman, Daniel. Thinking Fast and Slow. New York: Farrar, Straus, and Giroux, 2011. Hardcover.
Leonard Lee and Dan Ariely (2006), “Shopping Goals, Goal Concreteness, and Conditional Promotions.” Journal of Consumer Research Vol. 33: 60-70.
G. Lowenstein and J.S. Lerner(2003), "Role Of Affect In Decision Making." Handbook Of Affective Sciences, Vol. 31: 619-641.
J. Martin. (2017, Dec 22nd), “Chatbot Design Trends 2018.” Retrieved From: https://chatbotsmagazine.com/chatbot-design-trends-2018-253fb356d3a3
K.L. Milkman, T. Rogers, and M.H. Bazerman(2008), "Harnessing Our Inner Angels And Demons: What We Have Learned From Want/Should Conflicts And How That Knowledge Can Help Us Reduce
Short-Sighted Decision Making." Perspectives on Psychological Science, Vol. 3, No. 4, 324-338.
K.L. Milkman and J. Beshears (2009), "Mental Accounting and Small Windfalls: Evidence From An Online Grocer." Journal of Economic Behavior and Organization, Vol. 71, No. 2, 384-394.
Stanton et. al. (2014), "Effects Of Induced Moods On Economic Choices." Judgement And Decision Making, Vol. 9, No. 2: 167-175.
R.H. Thaler, S. Benartzi(2004), "Save More Tomorrow: Using Behavioral Economics To Increase Employee Saving." Journal Of Political Economy, Vol. 112, No. 1, pp. S164-S187.
Study Probes Spending Habits Of Millennial Grocery Shoppers(2016, May 24th). Retrieved From: https://progressivegrocer.com/study-probes-spending-habits-millennial-grocery-shoppers
Welcome everyone! It is an honor to be here today! How many of you are from the ECS school? Any from BBS? Any from ATEC?
I’m a proud UTD alumnus where I did my Bachelors and Masters in Cognitive Science and the Applied Cognition and Neuroscience program. These served me well as I work as a UX Researcher at BMC Software in Austin, TX where I examine the intersection between human behavior, design, and technology in solving enterprise design problems with one of the focus areas being on chatbots.
So just a show of hands, how many of you are familiar with what a chatbot is?
For the people, who have used a chatbot what was the experience like?
For the talk, I want to take you all back to a time when I worked on designing content for a chatbot called Olivia. Talk about the process involved in designing a chatbot and also how much has changed in the emerging field of conversational design.
So you might be wondering who exactly is Olivia?
Here’s the final form of what the app looks like.
Our main target audience for Olivia are people who are early career professionals who are getting adapted to being financially independent. That means paying for their own things like food, rent, travel, etc.
Now you maybe wondering, why is Olivia a Chatbot? To understand the decision to go the chatbot route, let’s take a look at this.
You may recognize these fintech apps as well as apps that are similar. What do you notice about these apps? When looking at your checking account what sort of feelings arise? What do you tend to do after looking at how much money you have in your account? They tell you how much money you have in your account, but not how you can best save money.
In other words, money is a behavioral problem, not a math problem. People are aware that they need to save money, but aren’t sure the best ways to. But what if there was a way they could get some suggestions?
With that we set out with some design goals to help the user save money, including providing tips in a variety of spend categories such as grocery shopping, eating out, clothes shopping, travel, etc. Making sure the personality was friendly and persuasive and not nagging like someone’s mother. And making sure the tips were actionable such as providing a link to the coupon app, or letting the user know when the nearest sales are, or even links to resources for most cost effective DIYs. So with that, let’s dig a little deeper into what went into the design of Olivia’s conversations with the use case of grocery shopping.
So with that, let’s dig a little deeper into the design process behind the conversations of Olivia with the use case of Grocery Shopping.
The first step is to understand Bill’s Grocery Shopping habits.
We conducted user interviews with people in our target demographic to understand how they go grocery shopping as well as some secondary research from credible articles to understand the scale of the problem. The information we got from this was then distilled into a journey map.
To help visualize the information we learned from the user interviews and secondary research, we created a journey map to make it easier to spot insights and create opportunities on what would help our users cut down on their grocery bill.
Common insights uncovered was a lack of a fixed day for grocery shopping, not creating a grocery shopping list, getting tempted by store items, and also determining whether to buy bulk or single pack or fresh or frozen to name a few.
Now that we have uncovered some problem areas it was now time to gather grocery shopping tips to help our users.
Getting grocery shopping tips involved conducting secondary research from various sources online and then aggregating them into a spreadsheet organized basically by the type of tip, a short gist of the tip, and the specific details of the tip and the source it was from.
After gathering the tips, it was time to plan out the conversational flow
Now a Conversation Flow is a way to outline how the user progresses through the conversation with Olivia. We had to think about Olivia would bring up grocery shopping spending activity, the types of choices presented to the user to progress through the conversation, and the order in which the tips are presented. Now these maps got really complex.
Once we settled on a conversational flow, it was now time to design the dialogue.
While the experience is designed around text, we wanted to make sure that the text was easy to read, engaging, insightful, actionable, and fit within the screen real estate space. Designing dialogue was a very iterative process with lots of revisions made till we were able to hit the sweet spot.
After various iterations on the dialogue, we were able to reach the final design that looks like this. Getting right level of friendliness, good amount of text and use of progressive disclosure.
So what were the results?
Well, when I last checked with my former manager the results are very promising as Olivia has been able to save 6-7% of user’s monthly income.
But most importantly, the app has a really high retention rate among fintech apps which is very crucial for conversational apps.
Now I have worked on Olivia from 2015-2016 and at that time a lot of my work involved learning and experimenting as I went. There weren’t may resources when it came to how to design conversational experience.
However, since then, a lot has changed in the conversational design space.
Before I would be using makeshift tools like mindmapping software like Draw.IO as well as Google Sheets to plan and create dialogues, but now there specific tools for designing chatbots without the need to code.
Additionally, new tools have released to measure chatbot analytics
There are dedicated roles for designing conversational experiences
And some companies like Capital One have dedicated divisions for Conversational Design
However, we are still at the beginning of this journey as we try to solidify design practices for chatbots and how to properly conduct usability tests. Hell, personality design is still a mysterious topic that needs to be explored.
So let’s be brave and embark on this journey! Thank You!