Guidient

AI learning from past interactions
Building on user past success
Sharing progress with support network
Simple chat interface
Priority and progress tracking

Inspiration

We saw the capability for AI to solve for more complex, social issues, and in combination with one of our team member having ADHD, and our collective interest in unique and contextual, we arrived at our solution.

What it does: We build informed, human support networks and tailored learning systems that grow alongside ADHD learners with the help of an AI guide.

To get the full story of Guidient, click here: https://www.figma.com/deck/NBuQaWH56UEMTzQd8BWGHf

How we built it

We built a web application that allows user to share their concerns and get response from the AI with user preference profile and their information(tasks, exams, etc). The frontend is in an interface of chatting it collects user's input and send it to the server with the previous information about the user collected from previous conversations and information shared by the user's support network (family, instructor, school). The AI is instructed to provide guidance to the user with all the data. The response from the server will then show on the chat box.

Challenges we ran into

Provide instruction to the API so it provides tailored guidance to the user.

Accomplishments that we're proud of

Running a user test during a hackathon and getting comprehensive feedback. Collecting 2 additional data points from a Google form survey to validate our product strategy and direction. Team meeting, discussing, refining, and flushing out the general structure of the finalized product within the first 6 hours of the hackathon

What we learned

Leveraging AI tools during the ideation and prototyping phase was essential to our rapid working pace. It is tricky to limit the scope of an ambitious idea, but it's doable with a well rounded but value-aligned team. There is no such thing as planning a timeline before the hackathon - shit will go off schedule. Everything will somehow work out at the end