AI in construction: A tender scanner for a construction company
With over 5,000 potential tenders and limited personnel resources, a leading international construction company faced a significant challenge: How can the tender processing be optimized so that resources can be concentrated on promising tenders? Using artificial intelligence, a solution was created that relieved the team, optimized resource utilization, increased the success rate, and enabled significant cost savings.
The importance of tenders in the construction industry
Tenders are official invitations to submit bids for construction projects, published by companies or public clients.
At any given time, there are numerous open tenders. Construction companies must analyze all these and identify promising projects. Due to limited resources, participation can only occur in tenders where they believe they can deliver good results.
Efficient processing and selection of the most promising tenders increase the chances of successful project wins and maximize resource utilization.
About the client
The client is a renowned international construction company with over a hundred years of tradition. With several thousand employees and hundreds of locations worldwide, the company has established itself as a leading player in the construction industry.
Industry: Construction
Project Year: 2023
Initial situation: the challenges of the construction industry
The client faced a significant challenge: managing and processing over 5,000 potential tenders at any given time. Due to the enormous volume, it had become impossible to allocate sufficient personnel resources to comprehensively review or process each tender.
An automated solution was needed: an AI-based system that could pre-select tenders based on the probability of winning, thus optimizing resource utilization and increasing efficiency.
Solution and development of the AI model
A custom AI model was developed to increase efficiency in tender processing. Sample data was reviewed, restructured, and cleaned.
Based on historical tender data and other internal information (whether the tender was won or lost, whether there were successful projects with the involved clients in the past, etc.), a neural network was developed. This network can individually predict the probability of success in participating in new tenders.
In addition to structured data such as contract volume, project duration, etc., free text from the tenders (e.g., project description) was also used for training and prediction. These contain valuable information for the probability of success. Such unstructured data cannot usually be processed by classification systems.
To vectorize the unstructured free text into numerical values, an embedding model from Aleph Alpha was used.
The resulting AI model classifies new tenders according to their probability of success with an accuracy of over 75%.
This automated analysis provides more accurate results than previous manual evaluations by the client’s business experts, with significantly reduced personnel effort.
Result: optimized resource utilization through AI in construction
The project successfully demonstrated the feasibility of using machine learning for the automatic pre-selection of tenders. This led to optimized resource utilization, higher success rates, and significant cost savings.
The client can now focus on the most promising opportunities, ensuring a more efficient and effective tender management process.
- Optimized Resource Utilization: By focusing on the most promising tenders, the company can better utilize its resources.
- Higher Success Rate: A targeted approach leads to significantly higher success rates in tender processing.
- Professional Implementation: The expertise of the ONTEC AI team ensured a highly professional implementation in an excellent working atmosphere.
- Cost Savings: Significant cost savings were achieved through outsourcing the development.
Conclusion and key takeaways: what AI can achieve in the construction industry
This case study with a leading international construction company shows how artificial intelligence and data analysis can significantly increase the efficiency and success rate of construction companies.
An AI-supported tender scanner was developed and implemented using machine learning and advanced data analysis.
Building on the developed model, the company was able to better analyze the vast amount of tenders and focus on the most promising ones.
- Optimized Resource Utilization: Focusing on promising tenders improves efficiency.
- Increased Success Rate: Targeted processing by AI increases the success rate in tenders.
- Cost Savings: Outsourcing and AI-supported processes significantly reduce costs.