Query Distribution System

Inspiration

In workspace, there are many tasks have to assign to different teams or colleagues, and it wastes lots of time. To address this problem, we aim to create a system to bring better workplace experience with multi-AI agent.

What it does

In the system, it can receive the customer's feedbacks and internal task assigning instructions, identifies the problem of those query and assign to the corresponding teams/colleagues to solve it. A summary report is also generated for users to have a better view on what query they are handling. On the other hand, the system is capable to arrange an optimised work schedule for the colleagues base on the task they are currently on.

How we built it

Each of our technical member first worked on the API handling (AI model and database) separately. We used Google Gemini as our AI model to do auto assigning and scheduling. For the database part, we used the sheets in Lark as our tasks database. After that, we combine each pieces into our final product.

Challenges we ran into

During the development, we found that it is challenging on decide what platform and model is suitable in our situation.

Accomplishments that we're proud of

We are very satisfied with our final product, as we are able to finish our system with all core function before deadline. Also, we are confident that our system can be used in workspace and optimised their experience.

What we learned

In this Hackathon, we learned lots of skills about handling API, from the decision on choosing which platform is suitable for our project, their limitations, to how to communicate with the platform with our codes. We believe that those skills are able to broaden our sight and make us a better system and AI developers.

What's next for our project

Although we have created all of the core functions for our system, there are many extensions that we can think of to bring the most seamless and optimised workspace experience to the users. For example, integrated more widely-used platform to make it more universal, and add further function like reassigning system. We hope to improve and make our project more complete in the future.