Pogo Airlines
Chatbot Design & Writing Case Study
Overview:
Pogo Airlines is a fictional airline I created a customer service-based chatbot for.
This project was completed as a final project for the UX Writers Collective Chatbot Writing & Design Course.
Timeline: 1 week
Role: Conversation Designer
Tools: Botmock, Lucidchart
Problem Statement:
Travelers need a fast and efficient way to manage itinerary changes and get answers to common questions.
Solution:
Implement a chatbot that allows travelers to make changes to their travel plans easily, give personalized travel recommendations, and answer common questions that many travelers have.
__________________________________________________________________________________________
Extensive and in-depth user research would be conducted if this were an actual project for a real company. I would interview millennials and ask what their experience has been with chatbots, what they liked and didn’t like, and what they would want most from the chatbot. I would also look at the context surrounding the mood and emotions when people are using the chatbot as well as location to influence the voice and tone.
Surveys would also be sent out to a broader audience to get a sense of what people want from the chatbot and what features should be implemented. I would then use affinity mapping to organize the results from the surveys and interviews, and create my key intents based on that.
I created a short brand document for the personality of the chatbot based on the information provided about Pogo Airlines and millennials:
I decided to go for a customer service-based chatbot for this particular project based on the information and research about the target users provided in the project brief. I believe there is a real need to fix these problems, as I have personally had many issues dealing with airline customer service throughout the years. I know many others have as well.
If the chatbot can actually help people solve their problems in an efficient way, Pogo Airlines will go a long way towards creating trust and loyalty for their brand, which can only be good for business.
Based on the information provided, I created 3 main intents:
Intent 1: Manage Reservation
Intent 2: Flight Status
Intent 3: Frequent Flier Information
I broke these intents down a little more and created a little sitemap to organize the main features of each intent, or goal:
I further expanded these intents by listing out what information needed to be asked and what information the bot could provide. Airline information is complicated, and it all needs to be mapped out very clearly so the bot is getting the information it needs to complete the tasks correctly and efficiently. Here is the expansion for managing a reservation:
Key Intent 1: Manage Reservation
Reservation Number - If user doesn’t have the reservation number, look up reservation by frequent flier number or transfer to customer service
Change Flight
-One way or roundtrip?
-New city of origin and destination city
-New return city of origin and destination city (Only for return trips)
-New date of travel
-New return date of travel (Only for roundtrip)
-Upgrade? (Default is economy)
-User chooses flight times if there is more than one
-Confirm all information
-Payment method (credit card, miles, travel credit)
-Confirm details and pay
-Send confirmation email
Rebook Flight
-Let the user know reason for flight cancellation or delayed flight
-If user is rebooked, ask if they want to keep the flight or choose alternative
-If alternative, list options and let user choose
-Give user option to keep original flight if they don’t like the alternative options
-Transfer to customer service if user doesn’t like any of the choices
-Send email confirmation if alternative flight is chosen
Cancel Flight
-Confirm cancellation
-Send email confirmation
After mapping out the essential information for the other intents, I started creating the script and flow chart. I started working on the script first, but I found it easier throughout the process to map out the conversation in the flow chart first and then fill out the script.
To see the complete script mapped out in Google Sheets, click here.
These are some of the basic flowcharts that I mapped out:
To see the complete flow chart, click here.
After I finished the flow chart, I used Botmock to create the prototype. Here is a demo for rebooking a flight:
Next Steps:
The next step is usability testing. This is a great way to figure out if there are any errors within the flows or if there is anything confusing within the wording that needs to be fixed. Further iteration and testing would occur after that.
Success also needs to be measured. Some ways to measure success include the change in the number of people using the chatbot, time spent on a task, rating for ease of use, completion rate for tasks, and drop-off rates. This tells the designer what is working and what isn't working, and where the chatbot can improve.
