Jonathan Dip, Ali Fakhreddine, Derek Chai, Saja Rashid, Tri Ha
Ghost kitchen owners often lack tools to effectively analyze their online order data. It is difficult to identify which menu items drive sales or how business decisions impact customer behavior. Restaurant teams need an efficient way to interpret delivery data and make fast, informed business decisions.
Our solution was to develop an AI-powered analytics tool integrated with the GoTab POS system. This system captures and analyzes online order data to understand item trends and customer data. It generates clear, actionable marketing recommendations tailored to restaurant data and empowers owners to optimize menus, improve targeting, and grow direct sales using their own data.
Over the 2 quarter project, our team utilized Agile sprints in 2 week intervals. We utilized the agile scrum process to manage our progression. We maintained weekly check-ins with our sponsor to refine scope, validate features and ensure alignment with what we were creating.
Our project can be split up into 6 phases -
1. requirements gathering, 2. low-fidelity wireframing, 3. high-fidelity prototyping, 4. GoTab integration, 5. testing and 6.deployment and final documentation
Design decisions were made by getting direct feedback from our sponsor, as well as collaborating as a team to figure out next steps. We had two main challenges, both with the third party API GoTab. In the first quarter, we didn't have access to the software until Week 9. In the second quarter, we had bugs and authentication/integration issues with GoTab's APIs. We dealt with these challenges in the second quarter by working on other elements of our design, such as the AI Chat Bot.
Below we have attached different types of documentation such as use case diagrams, user stories/personas, UML diagrams, user journey maps, low/hi fidelity wireframes, and class diagrams. In addition to this, we have our Figma designs, Trello, sprint reports, and our customer sided frontend.
We utilized many of the diagram tools to visually map all the interactions within the system in a way that helps everyone (including any non-technical users) understand what the system does and how users interact. Our user stories/personas helped to guide design decisions by capturing the goals of specific users and focus on their needs. Our Figma designs made it easy to create a prototype and get feedback from the sponsor before implementing into the final design. Trello, as well as our sprint reports, ensures that the team was aligned on deliverables, especially when working on different parts of the design.
Within our team, we had two methods for testing - manual testing and automated testing.
We manually tested the functional use of the "Order Now" data capture, the AI chatbot, as well as the live order reporting.
On top of this, we made sure the UI/UX was functional across the dashboard, and did regression testing after each push.
For automated testing, we made sure that the integration between the GoTab POS system and Order Now data was synched.
Outside of our team, we utilized User Testing/User Evaluation by giving our sponsor the opportunity to interact with our system. We found that we have the essentials of our product working, but we do have some accessibility issues. Users will struggle to navigate between pages with keyboard only, and users cannot interact with the AI Chat Bot using just the mouse. The combination of both a mouse and keyboard is required to maximize the use of our system, and the simple user interface makes it easy for a new restaurant owner to navigate and use.
Our final product includes an order alerts page, sales trends page, live orders dashboard with real-time order tracking, an AI chatbot for marketing insights, seamless backend integration with the GoTab system, and a simple and clear user interface that makes it easy for restaurant owners to use.
Order Alerts minimize delays and missed orders, improving accuracy and satisfaction.
Sales Trends show revenue patterns over time.
Live Orders reveal top marketing channels and customer behavior.
The AI Chatbot turns data into actionable recommendations to boost performance and sales.
In the first quarter, we were waiting for access to the GoTab API. Looking back, we would have been moving further along if we worked on other components while we waited for GoTab/our partner to give us access. In the second quarter, we were much more proactive about this, working on the AI Chatbot while we worked alongside the GoTab team to fix the API authentication issues we were having.
It was important for us to take initiative, rather than waiting for direction. This accelerated our progress and clarified our product goals. We encountered authentication issues when connecting to the GoTab API, which highlighted the importance of debugging. Additionally, we were overly optimistic on time estimates for certain features, which led to delays. By stepping back and having clearer task definitions and better team coordination, we improved sprint planning.