Understanding, extracting, translating: AI processing of construction tenders

©wirestock via freepik.com

Smart and error-free tenders

A European provider of infrastructure and environmental technology supports public and industrial clients in the planning and implementation of complex projects in the fields of water management, civil engineering and sustainable infrastructure solutions. With a team of specialised experts and modern machinery, the company stands for quality, reliability and technical excellence.

The tendering process: complex, time-consuming, indispensable

When we stepped in, the tendering process for our customers has been precise but very time-consuming. Together, we have developed a solution that revolutionises the entire process: automated processes standardise and translate Excel and PDF data in just a few minutes – ready for further processing in iTWO, the leading ERP software for construction projects and costing. The result: less effort, more efficiency and more freedom for what really matters: sustainable construction projects of the highest quality.

Created with Google Gemini

Optimising processes together

Large projects such as sewage treatment plants or pumping stations always start with a tender. For our customers, this means reviewing and processing extensive documentation. These usually consist of Excel files with itemised lists and lengthy PDF documents full of detailed descriptions. This presents two major challenges: different formats and a lack of standards.

Different formats and lack of standards

Tendering authorities, such as local authorities or engineering firms, work with their own formats. This makes further processing time-consuming, as PDFs in particular are often unstructured and non-standardised.

Crucial details in the long texts

The PDF documents contain essential descriptions that are indispensable for costing, such as specific requirements for materials or construction methods. This information is often missing from the Excel files and has to be laboriously extracted.

Overcoming language barriers

Many tenders are only available in French. However, these texts must be accurately translated into German for the calculation in order to avoid misunderstandings.

Intelligent processes for modern construction projects

Screenshot 1.1

To simplify the complex tendering process, we worked with our customer to develop a solution that combines technology and practicality. The result is a cloud-based application that meets modern requirements – intuitive, efficient and with intelligent use of AI.

A system that adapts to users

Usability has been a key focus right from the start. A clearly structured web interface ensures that complex tasks can be carried out with ease. Excel files and PDFs can be uploaded with just a few clicks, without the need for any technical knowledge. The application runs smoothly, whether in the office or on the go, thanks to modern cloud technology.

Structure data, generate GAEB files

The system automatically imports Excel specifications, processes them and creates standardised GAEB XML files. For our customers, this means that the data can be used directly in iTWO, the leading ERP software for the construction industry. This turns many individual parts into a structured whole.

When AI keeps track of everything

©immumi19 via freepik.com

Unstructured PDF documents posed the biggest challenge. These often contain detailed descriptions that are crucial for the calculation. This is where artificial intelligence comes into play: the texts are precisely read, analysed and assigned to the appropriate positions. At the same time, the system automatically translates French descriptions into German so that no details are lost.
ins Deutsche, sodass keine Details verloren gehen.

A tool for the future

This solution not only saves time, it also ensures greater security and reliability in the tendering process. What used to be tedious manual work can now be done in minutes – without losing track of the big picture. Our customer now has more freedom to concentrate on what matters most: sustainable, high-quality construction projects.

The benefits: more precise, faster, better

With the new solution, our customer has taken the tendering process to a whole new level. What used to be days of tedious manual work now takes just a few minutes – accurately and reliably. Sources of error have been minimised, and the time saved can now be devoted to core tasks: planning and calculating sustainable construction projects.

Digitization in the construction industry

©freepik.com

Our collaboration with our customers demonstrates the potential that digitalisation offers. In the future, the focus will continue to be on further optimising processes and making sensible use of new technologies. The solution is more than just a tool – it is a step towards the future of construction.

Key elements of the solution

The joint solution delivers impressive results:

  • Time savings
    Processing time has been reduced from weeks to just a few hours.
  • Greater precision
    AI ensures that no details are overlooked.
  • Uniform formats
    All data is GAEB-compliant and can be used directly.

Facts & Tech

Tech Stack: Python, FastAPI, GCP, Cloud Functions, Firebase, OpenRouter, Anthropic LLM Modelle

Picture of Oliver Tigges

Oliver Tigges

Oliver is one of the founders and managing directors of Inspired. In client projects, he is always particularly interested in mediating between the engine room and the boardroom. He also supports management in assessing the opportunities, risks and effects of pioneering and sustainable technology decisions.

Recent Posts

ein Handschlag zwischen zwei Personen
Press Releases

Inspired Consulting Becomes Official Eplan Technology Partner

Inspired Consulting joins the Eplan Partner Network, strengthening its role in maritime and electrical engineering software. With the upcoming launch of the Engineering Data Hub, the company introduces a powerful platform for streamlined collaboration, data integrity, and seamless integration across engineering tools.

Continue reading »