AIdriven optimering av anbudsprocessen - Effektivisering och automatisering

Sammanfattning

The construction industry is characterized by low productivity growth, where inefficient information flows, misinterpretations of tender documents, and evaluation errors often result in lost contracts, affecting the financial stability and growth of many companies. Compared to other sectors, the industry’s low level of digitalisation, combined with increasing societal demands (economic, environmental, and social) forces companies to invest significant time and resources into pricing accuracy and the formulation of qualitative descriptions. These qualitative factors, often referred to as soft parameters, such as sustainability, innovation, collaboration, and quality, have become increasingly important in public procurement processes. Digital technologies such as Artificial Intelligence (AI), particularly Generative AI tools like GPT models, have the potential to streamline tendering processes, including document interpretation, text generation, and price estimation. Despite promising results in early studies, the standardisation and practical application of AI tools within tender management remain limited. The aim of this study is to explore how Generative AI can enhance the handling of soft parameters in public tenders. The objective is to identify how AI can support construction companies in analysing large document sets, formulating execution plans, calculating total bid values, and communicating qualitative strengths more effectively. This, in turn, can increase the accuracy of tenders and the likelihood of winning contracts. The research addresses the following key questions: Q1: Which stages of the tendering process are the most time-consuming and challenging? Q2: Which soft parameters do public purchasers prioritise in tenders and how do they value these in the evaluation process? Q3: How can Generative AI practically improve these stages? Q4: Which parts of the tendering process benefit most from AI implementation? A qualitative research approach was adopted, combining literature review, interviews with contractors and clients, and a case study where various Generative AI tools were tested on real tender documents. This triangulation method provided both theoretical insights and practical evidence regarding the opportunities and limitations of AI in tendering. The interviews identified key challenges, while the case study demonstrated measurable benefits, showing how AI can support analysis, structure large volumes of text, identify relevant qualitative factors, and improve the clarity and consistency of tenders. Findings indicate that current tendering processes are largely manual, time-consuming, and highly dependent on individual expertise. The most resource-intensive stages are cost estimation, quantity take-offs, and text production. Respondents highlighted AI’s strong potential in text analysis, language enhancement, identification of key references, and integration into estimation systems, improving both efficiency and quality. Larger companies have already begun implementing AI-based systems to standardise and automate parts of the process using data from previous projects. The study identifies reference projects, previous experience, collaboration ability, and sustainability as the most critical soft parameters for winning bids. Public clients increasingly value these aspects, though evaluation often remains subjective and varies between organisations. Importantly, AI cannot replace human judgment and experience but can serve as a supportive tool that enables more strategic focus, better presentation of soft parameters, and higher-quality tenders. For AI to be successfully integrated, companies must structure and protect their data and train employees in secure and ethical AI use. Likewise, contracting authorities must formulate clear requirements for qualitative factors to ensure balanced evaluations beyond price. 3 In conclusion, the construction sector remains in an early and fragmented stage of digital transformation. Significant efforts, including standardised practices, clear policies, and specialised training, are needed to fully realise AI’s potential. The study concludes that Generative AI can play a key role in creating a more efficient, accurate, and sustainable tendering process and may become a strategic success factor for future competitiveness and growth in the construction industry.