CASE BRIEF & WORK

Having discovered a technological gap in the food industry, Byt Out decided to explore the untapped market and develop a new business model to handle the culinary needs of peculiar individuals with dietary preferences. Byt Out aims at reaching a vast majority of people who feel that their dietary preferences haven’t been given optimum attention by existing online food vending apps and website platforms. It is designed to help users personalise their dining experience across Europe and offer them the possibility of hand-picking restaurants or hubs within their location. At the core of their existence, they are on a mission to ensure that their food-service experience represents the needs, requirements, and lifestyle of the present society.

The Task at Hand

Byt Out was referred to us through a previous client in the UK who has now become one of our strategic partners. The client granted us access to their Functional Specification Documents (FSDs), which contained the mockup designs of the app as well as other information in both written and visual format. They precisely needed a Software Development Agency that understands the rudiments of Machine Learning and Artificial Intelligence and can integrate them into their app development. For the web development and machine learning bits, they required a company that can comfortably use Python programming stack.

How We Helped

Since we had successfully worked on Foodruns App in the past, we were ready to take up the challenge that came with working with Byt Out’s new business model. We created a simple, yet aesthetic landing page where people can indicate their interest to be a User or Outlet by pre-registering using their email address. It also has a blog section and an overview page where people can get in-depth knowledge of the brand and app. As requested by our client, we used Python programming stack for the Web Development and Machine Learning aspects, while we used Flutter for the Mobile App Development, and React for the front-end.

Apart from being a unique platform that provides the opportunity for people with dietary preferences, allergies, and special interests to choose from an array of Outlets available, our client had mapped out some other amazing features which we integrated into the Byt Out App. Some notable features of the App include:

• Machine Learning and Artificial Intelligence Integration: This is essentially used to study the profile, allergies, dietary preferences, and other special interests of Users on the platform. With the help of the ML models used, Users are given the most appropriate meal suggestions and shown the closest Outlets location to get served their meal of interest; remember, the brand is out to improve people’s dining experience. The App functions with these 4 Machine Learning models – Content-Based Filtering, Collaborative Filtering, Decision Tree, and Regression.

• Byt Profile includes subcategories like Byt Pals (similar to Followers), Recommended Byts, Liked Dishes, and Saved Outlets. Users can set their dietary needs (Allergies, Dietary, Don’t Like), profile, account, and even get Byt Support, which is a chat feature that allows the Support Team to respond instantly to a User’s message. Byt Pals is a community within the App; Pals are those who have similar dietary preferences as the User. The Recommended Byts is powered by Machine Learning models that recommend menus based on the information provided by the User in the Set Preferences section. It also makes it possible for Users to like menus they are interested in having or Outlets they’ll love to patronise.

• The Home screen displays ‘Urban on the go bite’ and ‘Top Byts near you.’ The Top Byts near you has two views – one shows the top meal recommendation, and the other is a map view that shows the Users, Outlets based on proximity. The subcategories under the Urban section show Outlets with Cuisine, which we broke down into Snacks, Street, and Barista.

• Byt Urban shows Hubs (collection of Outlets in a region). With this feature, you can find all Outlets within an area, view their address, available menus, and the number of people who like the menus at the Outlet.

• The Search Feature is quite different, in the sense that it automatically displays a User’s location. In essence, it can also be edited if you wish to search for meals in a different area. This feature is based on your registered preferences, location, and budget; you could also get suggestions relating to what you have inputted in the search bar.

Since the target Users of Byt Out are people with specific dietary preferences – vegetarians, pescatarians, vegans, as well as those with food allergies (fish, milk, nuts, casein, etc), we made provision for all Outlets on the App to upload their menu details including the ingredients used in preparing each meal. This is to enable the ML, and AI filter through each meal, check for ingredients that resemble a User’s allergies or dislikes and recommend only those that match their preferences. Furthermore, once an Outlet has uploaded its menu information (images, dietary categories, price, and availability status), it becomes visible on their dashboard. In contrast, their verification status, account information, and unique characteristics are only shown on the admin dashboard.

We began the project after a series of discovery meetings with 3 Software Engineers and a Project Manager on our team plus two members of the Byt Out team. Confluence and Jira were used for Project Management; the Software Development project was divided into eight sprints that spanned over four months and summed up to 70 MVPs-builds. Every two weeks, both teams (BFD and Byt Out teams) held an online video presentation using GoToMeeting or Zoom to discuss completed tasks as created in the backlog of issues on the Project Management Software. On the Jira platform, each development sprint had a minimum of 5 MVP-builds or issues (tasks), tickets (details about the issues), and sketches (pictorial representation of the issues).

Also, the Confluence pages created for the project included the Discovery Phase, Landing Page Structure, User Journey Flow, Wireframe FSD, Architecture, Server Configuration, and more. Aside from the Project Management Software used and scheduled Zoom meetings, the primary communication channel used was a WhatsApp group chat. Our breakthrough during this project was our ability to use Machine Learning to enhance Users’ chances of seeing only meal options according to their preferences without having to look through several images of alternatives they really don’t care about.

CASE RESULT