Leobit
AI-powered SaaS for CNC manufacturers
Developing a quoting SaaS application with artificial intelligence and sales process automation
ABOUT
the project
Client:
Location:
Switzerland
Company Size:
Industry:
Manufacturing
Solution:
Services:
The platform is a SaaS AI-based quoting app for CNC manufacturers who produce custom parts from metal or plastic using CNC machines. Equipped with a customizable dashboard tailored to each company’s specifics, the platform facilitates handling new requests and generating quotes. It also has a centralized and secure database where CNC manufacturers can efficiently store and retrieve information related to their CNC parts, 3D models, technical drawings, and previous quotes.
What sets this platform apart is its AI assistant, which learns from the data of each CNC manufacturer. This AI-driven feature automatically estimates machine running, set-up, and programming times for CNC parts. AI assistant streamlines and significantly enhances the quoting process, enabling CNC manufacturers to sell their products faster and more efficiently, from initial inquiry to the final offer.
Leobit demonstrates a diligent work ethic and takes full responsibility for the designated aspects of the project. Their expansive pool of experts is readily accessible whenever we require their expertise, making our collaboration particularly fruitful and dynamic.
CUSTOMER
The customer is a Swiss high-tech company specializing in Artificial Intelligence and Machine Learning, with deep mechanical engineering and contract manufacturing expertise. Their vision is to transform the manufacturing industry by adopting AI to automate the quotation process for CNC milled and turned parts.
BUSINESS
CHALLENGE
Our customer already had a primary platform that proved helpful for CNC manufacturers in calculating quotations. They wanted to grow their service and enhance it with new functionality but didn’t have enough skills and capabilities to do that in-house. Consequently, they opted to augment their internal team for development and testing.
In pursuit of this goal, the client chose Leobit to collaborate with their in-house development team. The joint effort aimed at the development and testing of new modules, features, and automation to enhance and grow their platform.
Why
Leobit
The customer discovered Leobit online and was notably impressed by the company's experience in .NET and custom software development, skill set, references, and projects in our portfolio. Seeking a team that could bring technical know-how to the project, Leobit emerged as a fitting choice. It's worth noting that Leobit demonstrated a forward-thinking approach and was ready to work with AI/ML even before it became a mainstream trend.
What initially began as a straightforward team augmentation has evolved into a substantial partnership. We have progressively taken on approximately two-thirds of the entire development and testing capabilities, making a considerable contribution to the success and ongoing evolution of the project.
Project
in detail
The full-stack developers at Leobit proficiently use Angular, .NET, and C# to craft modules for the platform. These modules facilitate various functionalities, including managing incoming quote requests, aiding users in calculating quotes, establishing pricing, and offering an online shop where buyers can accept or decline a quote. Additionally, Leobit has successfully integrated cloud storage into the platform, enhancing its overall capabilities.
Machine Learning Integration
Leobit was responsible for integrating ML-generated information into a web application. The client's software employs various ML algorithms connected to different libraries for analyzing and calculating CNC production time. Our team created services to retrieve, process, and transform information from ML services and invoke ML methods as needed.
We also developed a service responsible for indexing and constructing the entire ML model training history. This allows manufacturers to visualize how their ML model underwent training and influenced the results showcased on the website.
ML Model Processing and Storage
Another crucial aspect of the project was optimizing and storing manufacturers’ trained models. Once we obtain the results from the ML model, they undergo further processing and transformation for accurate display in the web application. Transformed data finds its home in MongoDB, complemented by data directly sourced from Azure Blob Storage.
We suggested minimizing ML model size for its swift loading and unloading into ML services. To do that, we developed a service that, based on the system's workflow, groups models together, archives them, and unloads them to the cloud. This allowed us to significantly boost the system's performance.
3D Modeling
To showcase CNC milled and turned parts in 3D format, we integrated the Three.js open-source library. It allowed us turning technical drawings users upload into dynamic 3D computer graphics directly in a web browser using WebGL. The 3D model dynamically adjusts in real-time based on material and other specifications users select for their detail.
Location-Based Directory
CNC manufacturers are presented to the customer based on proximity, a feature made possible by leveraging the Azure Geolocation service. This service retrieves the two-letter country/region code for an IP address, aiding in the seamless display and categorization of CNC manufacturers based on their geographical location.
Cloud Cost Optimization
Leobit’s experts thoroughly analyzed cloud usage and came up with several suggestions to improve project performance, optimize cloud resources and cut costs. Here’s what we achieved:
1.Shifting from a production to a development subscription to host all development and testing activities helped us cut down on expenses
2. Migrating the project from the Azure message broker to the Apache-based one and hosting it on a dedicated server helped us reduce costs and improve the software performance
3.Transitioning from one NoSQL database to another helped us cut expenses while expanding our utilization capabilities
4.Dockerizing our applications facilitated better management and cost reduction
Additionally, we introduced APIs for integration with client systems, replacing custom services we previously hosted for each integration.
The Journey
Behind Our Success
people in the team
CNC parts manufactured
clients serving with platform
years of partnership
Technology Solutions
-
90% code coverage and 95% coverage by automated test scripts integrated into the continuous delivery agile process.
-
90% cost reduction thanks to switching from the Gitlab version control system to Bitbucket.
-
Seamless communication between manufacturers and their clients thanks to integrated Sendgrid API.
-
Improved performance and cost savings thanks to migrating several app components from Azure to AWS.
-
GDPR compliance thanks to adopting robust data protection mechanisms.
Value
delivered
-
Unified platform offering a user-friendly interface for submitting requests, calculating quotes, setting pricing, and managing the entire quoting workflow efficiently.
-
More accurate and quick runtime estimation thanks to an automated evaluation of technical drawing.
-
Saving manufacturers up to 500 hours a year in the CNC quotation process thanks to a smart AI-based assistant.
-
Thanks to our cooperation, the customer is now serving more than 3,000 clients.
-
Our solution helped manufacturers estimate and manufacture more than 320,000 CNC parts.