We were inspired to create Emotion Pal based on a combination of our interests and expertise. We created Emotion Pal as an educational tool to highlight the important role emotions play in our lives. Our emotional experiences give us information about what’s going on and prepare us to take necessary action. Unfortunately, people are often not in touch with their feelings or can become overwhelmed by their feelings. The Alexa platform is perfectly suited for verbally guiding users toward a better understanding of their emotions, which is what Emotion Pal does. The more we can understand our emotions, the more we are equipped to make decisions towards a happy, fulfilling life.

What it does

Emotion Pal opens by prompting users to pick a word that best identifies what they’re feeling in that moment. Users can report a whole range of emotions, such as “scared”, “angry”, “guilty”, and “happy”, among many others. Emotion Pal responds to over 200 different emotion words.

Then, Emotion Pal offers two separate functionalities. In the first functionality, Emotion Pal will educate users about the different components of the emotion they’re feeling, including its triggers, how it’s felt in the body, how it’s expressed, and what the user might need. The content is based on current psychology research on emotions. The goal of this first part of Emotion Pal is to validate the user’s experience with emotion education. Sometimes, this knowledge is enough for people to know how to proceed. Other times, some extra support is needed to sort through what to do.

If users need extra support, they can purchase the second functionality of Emotion Pal, called “emotion coping tools”. This one-time purchase offers unlimited access to two types of coping tools called “check the facts” and “opposite action”, adapted from a clinical approach called Dialectical Behavior Therapy. Emotion Pal will guide users through using these tools. First, users are asked a series of questions to determine if their emotion is consistent with the objective facts of the situation. If so, then users are given ideas for how to act on their emotion in a productive way. If not, then users are given ideas for how to act opposite to their emotion instead. The goal of this second part of Emotion Pal is to offer concrete strategies for evaluating whether or not it is helpful to act on what the user is feeling.

How we built it

On the technical side, we built Emotion Pal using Voiceflow, which is a tool for designing and building Alexa skills. We based the content on psychology research on human emotions, as well as therapy tools from several evidence-based clinical approaches, including Dialectical Behavior Therapy and Emotion-Focused Therapy.


One of our biggest challenges was to make Alexa recognize over 200 emotion words and guide users down the correct conversational path from this input. Currently, Alexa does not have native “emotion” intents. Because this functionality was central to Emotion Pal, we had to build an emotion intent that recognized emotion words and a sorting function that provided the appropriate response based on user input.

Accomplishments that we're proud of

We are very proud of having created the first Alexa skill that integrates psychology research and therapy tools into a modality that is both user-friendly and technologically-advanced. Emotion Pal leverages the intimacy of conversation that is inherent in the Alexa interface, in order to validate how users are feelings with the ultimate goal of empowering them to create happier lives for themselves.

What we learned

Since this was our first skill to integrate in-skill purchase, we had to learn various technical aspects of enabling a one-time purchase (i.e., creating an ISP product, integrating the purchase option in the dialogue, testing the ISP in the developer’s console). We also wanted the free functionality of Emotion Pal to be available to markets without ISP capability (e.g., en-CA). We had to learn how to design a skill that restricts conversational paths to users outside en-US and en-GB, while still being valuable to those users. We also learned how to create custom intents for the variety of emotion words that users can report, and how to sort over 200 potential user inputs down the correct conversational path.

What's next for Emotion Pal

The next steps for Emotion Pal are to add more emotion words and add more therapy tools. We currently recognize over 200 emotions, but there are many other nuanced emotions not currently captured. We want to ensure users feel validated and supported, no matter how complicated their emotion. We also want to add other coping tools depending on users’ needs. Emotion Pal currently offers two coping tools (“check the facts” and “opposite action” adapted from Dialectical Behavior Therapy), but there are many other relevant tools used in therapy. Adding other types of coping tools will make Emotion Pal more helpful for users in a wider variety of situations.

Built With

  • voiceflow
Share this project: