Project Background

An Alexa Skill?

Alexa (the Amazon voice service behind millions of devices including the Amazon Echo) provides capabilities, or skills, that enable customers to create a more personalized experience. Amazon provides an Alexa Skills Kit (ASK) which is a collection of self-service APIs, tools, documentation, and code samples that enable developers to build and host most skills for free using Amazon Web Services.

About the Hackathon

General Assembly hosted an Alexa Hackathon where UX Designers and Web Developers had 8 hours to design an Alexa Skill. The theme of the hackathon was to develop an Alexa skill that would help enhance the user's wellbeing.

Teams had 30 minutes of brainstorming to pitch ideas and form teams, then 7.5 hours to build our skill. During initial planning, our team identified these additional requirements/contraints:

-Must be fully functioning in 8 hours
-No data mining/scraping available
-No database-building available


Pen & Paper, Whiteboard, Amazon Developer Console, AWS Management Console, Adobe Illustrator


Fully-deployable Alexa skill, branding and logo, final presentation with UX research summary


I completed user interviews and a survey to better understand users' current eating and exercise habits and the context in which they would be using CalPal. I also incorporated health promotion and motivational interviewing research to help guide the language CalPal would use when interacting with users.

Surveys and Interviews

I kept the research direct, quick, and no-cost. I led the development of a survey and an interview script. Participants were recruited from other Hackathon teams and I used my background in field/guerilla marketing to successfully recruit participants from the general public outside our building.

Secondary Research

Drawing on my professional experience in health promotion, I consulted articles on evidence-based motivational interviewing techniques used by health professionals when working with clients who are seeking to change health behaviors. I advocated for incorporating this into the language used by our skill.


Drawing information identified in my research, I worked with my team to create a realistic, evidence-based context scenario. I also analyzed the results of our interviews/survey to make recommendations for the developers on how our skill would interact with and respond to users.

Context Scenario

A majority of users reported that they would use this after an unhealthy eating choice was made (rather than before). These situations usually involved going out to eat and drink with friends or coworkers and consuming too much.

Users reported feeling the following in this scenario:
-guilt about their eating decision
-determination to do better
-desire to get back on track

Survey Results

Walking was the most common activity reported, followed by running, biking, and swimming. The survey showed that users preferred hearing time units (minutes, hours) rather than distance (miles, km) in our skill's responses.

Research-Informed Design

I did more user testing to understand the ways users might ask CalPal for information, then worked with the developers to make sure CalPal would respond in a clear and natural way.

Natural Language

Through brainstorming and user testing, I identified the possible phrases users might naturally use to ask CalPal for information. I created several utterances that CalPal would respond to that were congruent with this natural language and handed them off to our developers.

Error Prevention

Error prevention messages are especially important when interacting with voice recognition technology. Through user testing, I identified possible errors, then crafted responses for these error states that incorporated techniques from motivational interviewing research to minimize frustration and encourage users to achieve their goals.

User: “How long do I have to ice skate to burn 500 calories?”

CalPal (Before): “I’m not sure about that activity?”

CalPal (After): “I’m not sure. But you could burn 500 calories by running for 1 hour.”


I incorporated themes from user interviews and secondary health promotion research to develop a name that was friendly, supportive, and non-judgemental.

Research has shown that interactions with health professionals and counselors that promote acceptance, rapport-building, and positioning oneself as an equal partner (rather than as an authority figure) work best at facilitating behavior change (learn more here ). Incorporating these themes, I came up with the name CalPal: a non-judgemental "pal" rather than an authority figure.

I worked with our team's visual designer to create an icon for CalPal that reflected the user scenario and incorporated research themes (i.e. burning off calories after overeating with friends).

The Solution

CalPal won 1st place in the Hackathon and is now a fully deployable Alexa skill.

Enable CalPal

or say "Alexa, enable CalPal"

See more of my work: