It is an hour past midnight and you are still wide awake. After scrolling through various social media feeds, you decide you need to talk to someone. The question is, who is actually available for a chat right now? Then you remember, your one-night owl buddy who is always up for a conversation the moment you say, “hey, are you awake?” Lo and behold, your buddy is awake, and even though he is busy eating a late night snack of pizzas, he is ready for some banter. You chat for a bit about things like favorite episodes of Stranger Things, how poorly rested he is and that it is getting late, and that you have work tomorrow morning so you say “good night” and fall back to bed.
You may be wondering, who is this person? Well, just to clarify it isn’t a person, but rather a chatbot fittingly named Insomnobot 3000, designed and developed by the sleep products company, Casper. The purpose of this chatbot is to allow people who have trouble sleeping to have somebody to converse with for a bit and then nudge them to fall back asleep.
At the surface, the exchange between the user and Insomnobot 3000 seems very human-like with the way it does small talk about its favorite TV Shows or some amusing stories about its day at the office. However, all of this is cleverly scripted — everything from the times you can text it (only available between 11 PM-5 AM) to the domain knowledge it possesses about specific TV shows, the way it talks about sleep, and how it diverts the conversation from areas it doesn’t have any knowledge about through relatable small talk.
What makes this all work though is in how the content fits with the context and situational awareness of the user it is tailoring towards. A concept that cognitive psychologists term as our scripts and schemas.
If you were to look back in history, many psychologists throughout the Twentieth Century have always pondered about how people learn and form memories. Psychologists like Jean Piaget and Fredric Bartlett discovered that all the information we process gets stored as small building block concepts known as schemas. However, it wasn’t until the late 60s and 70s that Roger Schank, cognitive psychologist/artificial intelligence pioneer, explored the concept further with the idea that memories are formed by associating and sequencing different schemas together to form a cohesive script.
For instance, when you are at a restaurant you might notice objects like a menu, a table, napkins, plates, utensils, etc. as well as the people such as the waiter and staff. Each of these individual objects and people have characteristics and functions we associate with. A menu helps us look at the food items available at a restaurant, a waiter’s duty is to bring food, and the purpose of utensils such as a spoon, fork and knife are meant to help you with eating the food. These end up getting encoded into our memory as a schematic, so we always know who does what, what each object does, and how to use them. To form the big picture, each schema gets connected with one another in a specific sequence which ultimately culminates into Roger Schank’s classic restaurant script which roughly flows as follows:
1.Person enters a restaurant and looks for a waiter
2. Waiter spots the person and seats them at an available table
3. Person gets served water and is handed a menu
4. Person then looks at the menu to pick out what food items they want
5. Waiter comes back and asks the person what they would like to order
6. Person then provides the order to Waiter
7. Waiter goes back to the kitchen to provide the chefs the order
8. After some time the food arrives
9. Person uses utensils on the table to eat
10. Waiter asks for the bill
11. Person pays the waiter and then leaves
This script seems cut and dry, but it serves as the basis for how one would expect a conventional restaurant experience to unfold based on the prior experiences and associations made with our sensemaking minds.
With the understanding that humans have tendencies to build associations from context and prior knowledge, there are a lot of actionable items that Chatbot Builders can take away when creating a conversational experience that feels intuitive and personable.
Understand Your Target User’s Shoes (Or Slippers)
When creating an experience that is linear and one-to-one in nature, a rigid experience that doesn’t follow the user’s expected pattern of steps and actions can leave them frustrated. Rather than making assumptions about your users, conduct interviews, contextual inquiries, and usability tests with them to understand how they do things and what sorts of information are necessary to help them accomplish their tasks.
For instance, in an interview with Cnet, the designers and developers behind Insomnibot 3000 stated that through extensive usability testing at various times of the day, they were able to narrow down the time period for conversation between 11 PM — 5 PM as well as learn about the common topics people tended to discuss during those times.
To make the day in the life your target user more tangible consider creating an experience/journey map that captures the actions, context, and emotions of the user. Mapping it out helps you outline a conversational scenario that will make sense to the user.
Design The Dialogue With Your User’s Mannerisms And Colloquialisms In Mind
When designing the dialogue, keep in mind what you have learned about your user’s behavior. In the case of Insomnibot 3000, the learnings from user research helped craft the personality of the chatbot to be someone who also has trouble falling asleep and loves to binge on Pizza and TV shows like Seinfeld and Stranger Things.
The moment when users felt that the conversation felt human-like was the moment when the designers and developers realized that they were on the right track. People tend to build a liking to things that they can relate to, and by imbibing that in your dialogue, your users will feel like they are interacting with a human that understands them as opposed to a bot.
Building conversational experiences is as much a behavioral science problem as it is a development and design problem. It requires more than just engineering a smart algorithm and adding dialogue but also understanding the nuances of human behavior and communication to provide an experience that matches the user expectations of familiarity and authenticity.
In conclusion, I encourage Chatbot Builders out there to truly study the people they are solving the problem for in hopes they will make bots as useful and personable as the Insomnibot 3000.
Budiu, R.(2018). The User Experience Of Chatbots. Retrieved from https://www.nngroup.com/articles/chatbots/
Budiu, R., Whitenton, K. (2018). What Could An Intelligent Assistant Do For You? A Diary Study Of User Needs. Retrieved from https://www.nngroup.com/articles/intelligent-assistant-user-needs/
Erasmus, A.C., Boshoff, E., Rousseau, G.G.(2002). The Potential Of Using Script Theory In Consumer Behavior Research. Journal Of Family Ecology And Consumer Sciences, 30, 1–9.
Hall, E. (2018). Conversational Design. New York, NY: A Book Apart
Johnson, K.(2016). Why Mattress Startup Casper Built A Chatbot For Night Owls. Retrieved from https://venturebeat.com/2016/09/22/why-mattress-startup-casper-built-a-chatbot-for-night-owls/
Johnson-Laird, P.N. (1988). The Computer And The Mind — An Introduction To Cognitive Science. Cambridge, MA: Harvard University Press.
Katz, L. (2016). Insomniacs, This Chatbot Will Keep You Company. Retrieved from https://www.cnet.com/news/insomniacs-this-chatbot-will-keep-you-company/
Marr, B. (2019). 7+ Examples Of Online Chatbots And Virtual Digital Assistants In Practice. Retrieved from https://www.forbes.com/sites/bernardmarr/2019/02/11/7-amazing-examples-of-online-chatbots-and-virtual-digital-assistants-in-practice/#308e2b45533e
Minsky, M. (1986). The Society Of Mind. New York, NY: Simon & Schuster
Plant, K.L., Stanton, N.A.(2013). The Explanatory Power Of Schema Theory: Theoretical Foundations And Future Applications In Ergonomics. Ergonomics, 56, 1–15.
Ryan, K.J.(2016). Casper Moves Beyond Mattresses To… A Chatbot For Insomniacs. Retrieved from https://www.inc.com/kevin-j-ryan/casper-introduces-chatbot-for-insomniacs.html
Searle, J.R. (1981). Minds, Brains, And Programs. In Hoftstadter, D.R. The Mind’s I (353–373). New York, NY: Basic Books.