Co:nsole

💡INSPIRATION

Building a more empathetic world. Whether you are a healthcare worker, teacher, or corporate manager—empathy carries beyond the workplace or healthcare setting. We are first friends, parents, sons and daughters, sisters and brothers—and human. Our seemingly interconnected world continues to showcase declines in empathy due to crises in the world (racism, COVID-19, and economic insecurity) and increased isolation powered by technology. We were inspired by the emotional sensitivity displayed by the people in our lives and those working in positions that require that support for the mental well-being of others. We believe that empathy should be championed no matter your role. Knowing how to respond to difficult situations and sensitive discussions will pave the way for more kindness in our communities. With the goal of creating more empathetic interactions, we created Co:nsole.

📞WHAT IT DOES

Co:nsole is a way for everyday people to practice being there for other people. By chatting with our AI, Co:nsole trains core empathy skills such as listening, seeing the issue from others’ perspective and how to support others in times of need. You can select the situation and your role, whether you would be a helpline emergency worker, friend or Human Resources employee and get access to different chatting scenarios every time. While chatting, you can gauge how well you are doing based on our changing background, the sunrise getting warming as you continue to provide better responses and cooler when you make a mistake.

💻HOW WE BUILT THE PROJECT

Con:sole was building using many technologies, including a website front end with HTML, CSS and JavaScript, as well as a back end with Co:here’s generate and sentimental analysis API.

🌟CHALLENGES

This was our first time working with large language models and while Co:here’s API was simple to get started, it was challenging to generate the text we needed for our application. With the aim to use the API for conversation, we experimented with different generation models to determine which was the best fit for our use case. One of the best ways was to find connections between what we were building and past projects. We also encountered the limitations of NLP and recognized gaps between computer and human understanding. For example, we had to carefully craft prompts that would fit the criteria we were looking for in the generated text. Another challenge we faced was figuring out how we would

🏆 ACCOMPLISHMENTS THAT WE'RE PROUD OF

Working on this hack was a great learning experience for every single member of our team. We got to ideate and work on a product that fits our values. We also got to work on our front-end development skills and learned how to apply the Co:here API, which barely any of our team had experience with APIs before.

❓WHAT WE LEARNED

We learned that working together is an important part of the technical development process. Many of our team had a variety of skills, some were experienced in the front end, others had great product designing skills and some were most interested in the backend and working with Co:here. Through divide and conquer we were able to each learn what we wanted to learn the most and also create a great project we were all proud of.

👣WHAT'S NEXT

What do we have on the roadmap? We hope to leverage the expertise of mental health professionals and fine-tune our model and scoring of the responses users give. We also plan on improving the project features to include a dashboard and database storing the data on how much practice they’ve completed, average scores and commonly used phrases so users can gain insight into how to improve. Overall, Co:nsole has the capacity to build a more empathetic world, one footstep at a time.