The tech company saw opportunities in AI for improving operations and pricing accuracy. They collaborated with 3Cloud to develop a process and platform using Azure and open-source technologies. Azure DSVMs were used for memory-intensive development. NLP approaches for incorporating the large amount of text data available in written construction scopes and line-item descriptions were tested to analyze construction data stored in Azure SQL Databases and Azure Blob Storage. To craft the Proposal Review Assistant, 3Cloud architected a system of machine learning models and analytic modules to provide recommendations to Account Managers on which actions should be taken during the review process – with AI adoption and explainability the driving force behind AI development.
To solve Data Research challenges 3Cloud boiled down the the data research task into an easily digestible question, and landed on a deep learning similarity model to match supplier products and materials together based on their text descriptions, using the large history of research as labeled data. They tested the MVP model and gathered feedback from the research team before moving to production.
3Cloud’s AI solutions have had a significant impact on the technology company. They have set the business apart in the construction industry, showcasing their commitment to innovative construction solutions and efficiency. These AI solutions have also allowed the Data Research team to focus on more complex client challenges.
3Cloud’s AI solutions incorporate best practices in software development and data science, laying the groundwork for future AI product development. The Proposal Review Assistant has saved Account Managers hundreds of hours, leading to quicker and more accurate proposal reviews, empowering them to communicate effectively with clients.
The Data Research matching model MVP has shown potential for 20-40% time savings in research, with enthusiastic support from researchers. These solutions serve as the foundation for the company’s AI product development scaling