How Skoop tapped into a $12 million industry with NEOITO

Read how Skoop revolutionized food industry billing with custom-built AI and ML algorithms

Industry

Restaurant Franchise

Country

Project Overview

There is a serious lack of monitoring of the right portion and ingredients of food in restaurant businesses. Skoop was designed to solve this problem using IoT and ensure portion control for all orders.

Our task

AI Analytics

Image Detection

Hardware Development

01

Backstory

Travis, a young entrepreneur, uncovered a $12 million industry while working at a Chipotle restaurant. He found a discrepancy between add-ons listed on the bills of orders and what was actually added.

Travis came up with a solution to build software designed to identify the individual ingredients in a burrito or burrito bowl by analyzing images captured by a camera. The software can distinguish between different types of ingredients, added to the burrito bowl and accurately label them.

That’s when he realized a need for a tech partner with experience in handling advanced tech stacks like AI, ML, computer vision, and custom development. Travis approached NeoITO, impressed by our experience in integrating these advanced technologies. Thus began the journey of Skoop.

02

The need for Skoop

At the early stage of the project discussion, we realized the need for high-performing object detection hardware to create the best-performing data algorithm. We developed all the components from scratch from hardware, software, and design.

Skoop is made to cover three crucial needs:

Using AI analytics vision module to find the ingredients used.
Collect the images from the camera and convert them into usable data.
Process the data with the help of the backend server.

The ultimate aim was to build a product using computer vision to recognize the items on a serving table and monitor the hand movements connected to those objects.

Most assembly cycles are built on the Yolo V5-Deep shot algorithm, but it is ineffective in the Skoop case. So we build the algorithm from scratch.
Finding the appropriate hardware which can capture high-definition data without pixelation
Building a custom algorithm to track the ingredients
Develop a backend server to process data and an app that makes the data useful by adding the exact ingredients to the bill.

03

The Challenge

We needed high-quality data to build a high-performing object detection model. Upon receiving the video from the client, it became apparent that a quality enhancement was necessary for further processing.

When Travis found it challenging to fetch videos, we stepped forward and created the object detection model in our office and recorded videos for the project.
Tracking the hand – where the hand goes and which ingredient the hand takes. Analyzing the moment and tracking the number of scoops taken needed a custom model dataset.
As our hardware usage increased, the need to transition between software became pressing. Our tech team promptly adapted and excelled, ensuring optimal results.

04

The Neoito impact

With our custom-made database, any food chain can automize their extra add-on billing, and customers can easily track the ingredients from the app. The basic MVP of the project has been attained, and the project is in the next stage to become a full-fledged product. We continue to be their trusted technology partners.

NEOITO’s all-round support in developing Skoop was impactful in our product’s early success. We were able to test out the product on our customers ahead of time and get valuable feedback on features to make Skoop even better.

Travis

CEO, Skoop

05

The Result

Our expert tech team, along with more feedback from the founder after POC, achieved the basic MVP within 5 months. - We created a custom-made quadrant detection and tracking algorithm to track burritos along with hand movements. - Our custom-made algorithm can easily track more than 100 items simultaneously. - We built an app with Firebase for cloud storage. - Provided the end-to-end solution - Hardware, Design, and Software.