Disney Guest Services
Self Initiated project
The following was my design process developing the Disney Guest Services chatbot design proposal. There’s so much opportunity for Disney chatbots to give the same excellent service and magic that guests experience with real humans!
Research
I needed to understand what the best role of the chatbot should be: considering it’s Disney, and considering its guest support in a variety of areas: Parks & Resorts, Shop Disney, Star Wars, Games and apps and anything in between.
I developed a user persona, understanding the audience’s state of mind, goals, pains, preferences…etc
I researched chatbot design best practices, common mistakes
Examples of great chatbot designs from other companies were inspiring, such as Hello Fresh or Slack
Understanding that there are B2B companies that have chatbot technology to speed up production, such as Instabot, was important to know.
Goals & Deliverables
Project is designed to provide insights and guidance for the chatbot design
I want to collaborate with team members to ensure buy-in, and also improve the design
UX goals:
Must remain obvious to user that chatbot is not a real human
Must actually save user time to get answers they need
Must improve overall perception of Disney service
Must have the Disney magic touch
Deliverables: documentation showing research sources and key findings, user flows of ‘golden path’ to getting answers from chatbot as well as anticipated worst-case scenarios, all in an interactive prototype.
Personas:
Chatbot Personality
I first concepted generic bot designs, inspired with android bots and various research. Then later decided to go with the friendly Baymax—he is an actual Disney character who embodied the goals of what the chatbot needed to do. The opportunity to customize the character according to the product could be interesting. For example, on the Star Wars help site, could C3PO be your helpbot?
Team Brainstorm
I worked with the development team, but quite a bit with the team’s Product Specialist. We discussed what her experience is with common questions, anecdotal feedback from guests, and BI data on support email/live chat volume. We then worked through expectations, what team already knows, and blue sky ideas.
User Flows
I considered the ‘golden path’ of a user interacting with the chatbot. I also mocked up scenarios that were less ideal, and perhaps the most disastrous possibilities and how we would deal with them. Primary output: Mobile layout based on BI data of most common device. Secondary design considerations: desktop layout. Examples of user flow scenarios:
Easy resolve answers: user either picked from FAQ topics that chatbot suggested, or typed meaningful key words
User’s custom message not recognized
User is not getting the answer they wanted; escalate to real human
Collect quick ‘was this helpful’? feedback
User entered profanity
Prototype and Executive Presentation
Prototyping the idea helps to validate concepts on the flow before costly development has to start. I provided the team with an interactive prototype to show how the helpbot would initiate itself, and how the user would be expected to interact with it. This was then prepared to be presented to executive leadership (along with other documentation) to get the project approved!
Prototype
Please note: this was only a quick and fast preview of what chat might look like to launch and exit. Much of the conversation around the bot interactions themselves didn’t need to be prototyped yet with where we were at in the project.