A.L.F.R.E.D.: Automated Logistics & Flexible Rescheduling with Enhanced Dashboards

Inspiration

Modern businesses, especially those in retail and service industries, face the challenge of efficiently managing shifts and employee schedules. A.L.F.R.E.D. was born out of a desire to streamline this process using advanced technologies and APIs. By integrating Square's Team and Inventory API and Google's PaLM API, we aim to facilitate easy shift rescheduling and assist managers in planning shifts effectively.

What it does

A.L.F.R.E.D. is a comprehensive solution that offers two main interfaces: a chatbot interface and a schedule interface. The chatbot interface allows employees to request shift changes, and it suggests available employees based on their preferences and schedule. On the other hand, the schedule interface empowers managers to plan and visualize future shifts effortlessly. The system automatically sends a message and email to employees regarding rescheduling of their shifts, keeping them up to date and simplifying the process. It also has a page with all current requests, allowing managers to accept / deny rescheduling requests. Through intelligent integration with Square's Team and Inventory API, the system ensures that each shift is adequately staffed with capable and available employees. It further combined PaLM and Square's API to find the employee best suited to fill in for a shift.

How we built it

To create A.L.F.R.E.D., we leveraged Square's Team and Inventory API to access employee data and availability. Additionally, we integrated Google's Palm API to provide a chat-interface for employees and managers. We utilized a Next.Js frontend with TailwindCSS, with a python-flask backend supported by a PaLM's API integrated through LangChain.

Challenges we ran into

Integrating multiple APIs and ensuring seamless communication between the chatbot and schedule interfaces presented a significant challenge. It was difficult to reduce hallucinations for the LLM, but we surprisingly solved that by stumbling upon a research paper that outlined how to prompt-engineer different Large-Language-Models. Furthermore, integrating the backend and frontend proved difficult, but utilizing TypeScript helped reduce errors. We further interviewed over 30 employees across 5 firms to understand their struggles with onboarding and rescheduling, which helped us fine-tune the product and better understand the problem.

Accomplishments that we're proud of

We are proud to have successfully integrated Square's Team and Inventory API and Google's Palm API into a cohesive and user-friendly solution. The chatbot and schedule interfaces provide a seamless experience for both employees and managers, improving shift management efficiency and employee satisfaction.

What we learned

Throughout the development of A.L.F.R.E.D., we gained invaluable insights into API integration, real-time data processing, and user experience design using Next.js and TypeScript. We deepened our understanding of machine learning applications, particularly in the context of employee scheduling and availability prediction, leveraging Google's Palm API for enhanced functionality and accuracy.

What's next for A.L.F.R.E.D.

In the future, we plan to enhance A.L.F.R.E.D. by incorporating PaLM further improve employee questions and rescheduling. By acting as a membrane between employees and managers, we aim to increase productivity for both. Furthermore, we plan to further interview employees across firms to enhance our understanding of the problem and the subsequent solution we build.

Built With

Share this project:

Updates