01 The Challenge
A Useful Voice UI Cooking Skill for Amazon Alexa
While exploring Alexa skills, my research revealed that most users find a purely voice-driven cooking experience to be cumbersome at best. It can be difficult to follow complex recipes, and asking Alexa to constantly repeat herself becomes tedious quickly.
Using Alexa to browse for recipes by voice alone also has disadvantages versus more traditional means. Stalking through enticing photos of our meals-to-be produces a joy rivaled only by partaking in the meal itself, an experience which voice simply cannot replicate.
Enter: The Smoothie
Smoothies are simple. They require one cooking tool, no fiddling with temperatures, and the audience who is most likely to use an Alexa to help them cook is also the type most likely to want a quick, easy, and healthy meal with minimum chewing. I'm making an assumption about the chewing bit, but smoothies have the distinction of being largely uniform in appearance aside from color, bypassing another pain point other meal types encounter on VUI platforms.
My goal is to create a truly useful cooking skill which is not only suitable for voice UI, but augmented by it.
02 User Research
I performed a series of brief user interviews, both of users familiar with voice assistants and those who had only limited experience.
Questions focused on:
● Pain points users encountered while using VUI or traditional means to browse and follow recipes.
● Especially useful or delightful VUI features users have encountered, and exploring how they could be applied to a cooking experience.
● Exploring the habits users have while preparing meals in general, especially quick, unplanned ones.
● Users who had used VUI for cooking in the past found it was only useful for simple recipes.
● Users would like to be able to browse recipes by ingredients, if they are making a quick meal, they are only going to use what they have on hand.
● When planning future meals, users place a high priority on photos as it is a simple way to sort what looks interesting or not.
"Trying to get it to tell me an entire dish would be pretty obnoxious. Even five steps gets tedious."
"A 'random' recipe in that case is only useful if it doesn't send me to the store."
"There should be an option at the beginning that lets you permanently exclude things you don't like."
● Allow users to browse recipes by ingredients or categories (e.g. breakfast, healthy, fruit, etc.).
● Provide an adequate number of smoothie choices for the user to get value from the app.
● Make it easy for users to select preferred smoothies.
I recorded user quotes and observations onto post-it notes, grouping them together to create an affinity map. This would allow me to visualize common pain points and desired features. I also recorded common traits among my interviewees, which would assist in the creation of my user and system personas.
My user persona, Jared, is in his 30s is a self-described foodie, works in the tech industry, and recently received a blender as a gift. This persona provides enough information for an initial MVP, and I can create more should the need arise.
Belinda has her own persona, which will help me or any developers better develop her dialogue.
04 User Stories
I created user and job stories based off of my persona. I further developed some with sample dialogue which will help me plan my Intents, the strings of logic that make up individual features in an Alexa skill.
05 User Flows
I am currently refining my user flows, which provide the information architecture foundations needed to plan my script, Alexa intents, and Alexa slots.
06 Next Steps
This project is an active work in progress, with the end goal being a certified Alexa skill that can be downloaded for home use!
Next to do:
● Further refine and clean up user flows.
● Create skill script with sample utterances.
● Finish recipe list.
● Upload intent utterances into Alexa Developer Console.
● Upload intent script and other backend coding into AWS Lambda.
● Submit skill for certification via Amazon.
Thanks for looking!