— PROJECT NAME
Alexa Ice Cream Social
— ROLE
Customer Journey Mapping
Voice Design
— DATE
2021
— TEAM
VUI Designers
(for peer review)
For the past year, I have been a UX Designer with Alexa Shopping. I've enjoyed designing and collaborating on projects for our Echo Show and Echo Dot devices.
These projects are currently confidential. 😬
However, I can share a fun concept project I worked on, which envisions the potential of Alexa as an ice cream ordering assistant.
The Collaboration:
We chose Molly Moon’s as our potential partner for this ice cream endeavor.
From kitchen to cone, Molly Moon’s ice cream is the result of fine ingredients, happy staff, and local partnerships. The Seattle community is connected when people share sweet treats with the people they love. Now, they’re making ice cream orders easier by pairing with another homegrown favorite: Amazon’s Alexa.
The Problem:
Molly Moon’s Ice Cream would like to allow brick-and-mortar customers to use Alexa to order ahead, reducing the amount of time store associates spend taking orders, and letting customers who order this way to skip the line.
Design Challenges:
- • Design an experience on Alexa that allows a customer to order one or more menu items, with customizations, and be able to checkout and pay.
- • Design a VUI (Voice User Interface) experience that leverages the functionality enabled by Alexa Conversations.
Alexa Conversations:
Alexa Conversations (AC) is a set of “plug and play” VUI capabilities built by the Alexa org, for Alexa Skills, that helps maintain context and conversation state across multiple turns between the user and Alexa.
The Process
Customer Journey Map
What steps will users go through?
- I began the VUI process with a customer journey map to understand what steps users will take to place their order, and what their goals will be at each stage. I designed the customer journey for a new customer. For this user, introduction and orientation steps are important so they can aclamate and understand how to make choices with their personal assistant.
You can use this layout block to write as much or little as you'd like.
You may also paste in embed codes from popular services like Vimeo, SketchFab, and YouTube.
Interaction Flow
How will the user interact with Alexa?
From here, I began an interaction flow to help visualize the voice-driven experience. I considered edge cases, errors, and other responses off of the happy path.
This stage focuses on users possible actions, focusing on the happy path, while also considering some common edge and error cases.
The nature of ice cream ordering complicates the action phase, where users will be making choices for their ice cream order. Users must choose between order type (cup, waffle cone, sundae, or ice cream flight), order size (up to 3 scoops), flavors (15 choices), and toppings (12 choices). If users were to give their orders in a single turn (i.e., I would like 2 scoops of lavender honey ice cream in a cone with whipped cream topping), this results in millions of different possible user utterance for Alexa’s NLU to understand.
To increase the likelihood of Alexa understanding users’ orders, I broke down the order/action stage into 4 conversational turns: item selection, size selection, flavor selection, and topping selection.
Sample Dialogs, Errors, & Recovery
At this stage, I was ready to craft sample dialogs to design how the user and Alexa would get through turns in the interaction. I focused on a hero path: an Alexa user who is at home and using this skill for the first time. Because this is an interaction with many steps and turns, the conversation will have an open mic, meaning users will not need to say the wake word (Alexa) at every step.
While I focused on the “happy path,” the nature of VUI interactions necessitates accountability for users who off of this path. So, as the designer, I anticipated the most common alternatives users may select, and possible errors they could face in the process. These alternatives are based on Alexa Conversation capabilities, including the following:
These alternatives required alternative dialogs.
