Artificial intelligence project will reduce cost estimation to 60 seconds

Scientists at Boğaziçi University the Institute for Data Science and Artificial Intelligence have started work on the "Artificial Intelligence Assisted Cost Estimation Software Development Project from Technical Drawings". With the software to be developed, companies will be able to calculate the cost of the products they manufacture within 60 seconds based on the technical drawing using artificial intelligence algorithms.
Osman Yalçın

Researchers at Boğaziçi University's Institute for Data Science and Artificial Intelligence have initiated the "Artificial Intelligence Assisted Cost Estimation Software Development Project from Technical Drawings" The upcoming software aims to empower companies by enabling them to swiftly determine product manufacturing costs within just 60 seconds, utilizing artificial intelligence algorithms and technical drawings.

Led by Assoc. Prof. Dr. Şener Özönder and Dr. Hüseyin Oktay Altun, "Artificial Intelligence Assisted Cost Estimation Software Development Project from Technical Drawings' ' is underway at the Institute of Data Science and Artificial Intelligence, Boğaziçi University. Operating under the guidance of Assoc. Prof. Dr. Berk Ayvaz from Istanbul Commerce University, the project's primary goal is to streamline the cost estimation process and establish an effective pricing policy for products planned by companies, leveraging the capabilities of machine learning.

"THE FINANCIAL LOSS RISK WILL DECREASE."

The initiative, backed by TÜBİTAK 1711 Artificial Intelligence Ecosystem Call—an esteemed component of TÜBİTAK's competitive programs—will last 18 months. Dr..Hüseyin Oktay Altun, elucidates the inception process of the project idea as follows.

"When companies establish pricing for their current and upcoming products, this procedure often proves time-consuming, potentially leading to both financial and temporal setbacks. Moreover, inaccurate cost analyses underlying price quotations can further jeopardize companies financially. Our newly developed software delegates this responsibility to machine learning, offering a solution to reduce the risks associated with financial and time losses."


(Example Cost Analysis Within the Scope of the Project)

Dr. Altun, who expressed that artificial intelligence algorithms will analyze all relevant data of the company for pricing the software developed within the project, states, "The software will automate the cost estimation the company aims to make for its new or already existing product by scrutinizing the technical drawings of the products in the company. Thanks to machine learning, we will witness the continuous improvement of more accurate business processes each day. The resulting output will not only guide companies but also significantly expedite decision-making processes."

"THE PROJECT WILL STRENGTHEN THE ARTIFICIAL INTELLIGENCE ECOSYSTEM IN TURKEY"

Assistant Professor Dr. Şener Özönder from Boğaziçi University's Institute of Data Science and Artificial Intelligence highlights the crucial role of TÜBİTAK 1711 Artificial Intelligence Ecosystem Call in fostering both the development of products by entrepreneurs in the deep technology sector in Turkey and addressing the need for effective artificial intelligence-based solutions in both public and private sectors.


(Left to right in the photo: Assoc. Prof. Dr. Şener Özönder, Boğaziçi University Rector Prof. Dr. Mehmet Naci İnci, Minister of Industry and Technology Mehmet Fatih Kacır, TÜBİTAK President Prof. Dr. Hasan Mandal, Teknorot Otomotiv General Manager Cevat Aslan)

Assistant Professor Dr. Özönder comments on the collaboration established for the development of machine learning-supported software:

"This initiative is a university-industry collaboration project involving stakeholders, including the Institute of Data Science and Artificial Intelligence, where I am also a faculty member. The initiative, ArtificaX Informatics, Teknorot, a manufacturer of vehicle steering and suspension components, will be the customer of the software we have developed. In this project, we will develop a multimodal machine learning algorithm capable of simultaneously processing geometry, text, and part images from technical drawings used in the production phase and cost estimation. Embedded in the software, this algorithm will extract the technical drawing of the product from the company's ERP system. The artificial intelligence algorithm will predict quantities such as cost estimation, energy consumption, production time, and carbon emissions, conveying the results back to the ERP system. This will reduce cost estimation analyses, which typically take up to 10 days when done manually, to just 60 seconds. The goal here is to both reduce the company's costs in these processes and enhance global competitiveness by enabling rapid cost estimation and customer proposal generation. The developed software will bring benefits to every industry using technical drawings in production, from automotive to aviation."