RGF Staffing relieves clients’ with AI-powered collective agreement checker  

New regulations on equal pay for temporary workers created a significant administrative challenge for both the staffing sector and its clients in 2026. To address this, RGF Staffing the Netherlands partnered with Conclusion to develop the Collective Agreement Checker: an AI-supported solution that reduces administrative effort, saves considerable time and gives clients fast, clear insight into the correct remuneration of their temporary workforce.

 

From 1 January 2026, temporary workers in the Netherlands are entitled to equal remuneration. They must receive at least the same pay as permanent employees performing a comparable role within the client organisation. This brought important changes. Clients needed to ensure that collective labour agreements, employment conditions and company-specific policies were also correctly applied to temporary workers.

Adam Hassan, Director IT Technology & Business Operations at RGF Staffing the Netherlands, part of Recruit Holdings, one of the world's largest specialists in flexible workforce solutions and HR technology, recognised early on that the impact would be substantial. Not least because of the approaching 1 January 2026 deadline.

 

"We wanted to find a solution that would ease the burden for both RGF and our clients," says Adam. It soon became clear just how significant the operational challenge was. "Within a short timeframe, we had to translate a large volume of collective agreements and company-specific policies into clear insight on the correct remuneration for our temporary workforce. That was a major undertaking."

The challenge involved hundreds of collective agreements and company-specific policies, each ranging from 60 to 200 pages. Adam continues: "No two documents were structured in the same way and they were submitted in many different formats. Some were Word files, others were poorly scanned PDFs. Essential information was often contained in images that were difficult to read. To bring order to the chaos, our CEO, Maximilian Krijgsman, gave us the support and freedom to explore an automated solution. That is when we approached Conclusion, a partner we already knew from previous collaborations."

For Conclusion project manager Kelly Meijers, the challenge presented more than just a compliance issue. "The legislative change created a unique opportunity to redesign the process and make it significantly more efficient. We challenged RGF by asking: why simply automate the existing process? Let's redesign it and explore how AI can support it in the most effective way."

 

To map the possibilities, teams from RGF Staffing and Conclusion organised an event storming session in July 2025. This interactive workshop analysed the existing process and identified opportunities for improvement. Not only collective agreement specialists participated, but also stakeholders from across the organisation, including end users.

Nick Radunovic, AI Engineer at Conclusion, explains: "Following the event storming session, the initial idea was to implement an off-the-shelf AI solution. In practice, this approach proved too slow and insufficiently scalable for the vast number of documents, data points and information components involved. We quickly shifted towards building a solution capable of analysing multiple collective agreements and company policies and presenting the results in a structured way."

 

Adam adds: "You cannot apply a standard solution to a non-standardised process, not even with AI. That prompted us to take a closer look at how documents were submitted, processed and reviewed. We soon concluded that an off-the-shelf AI solution was not realistic." Instead, the team chose to build an application that automatically processes collective agreements and company policies, powered by a robust AI model and supported by a human-in-the-loop review process. This became the RGF Staffing Collective Agreement Checker.

Adam explains how the solution works: "Users upload the client's collective agreement and supporting policies. At the click of a button, the Collective Agreement Checker reads the documents, extracts the relevant information and translates the agreement components into an equal remuneration form."

 

Following an initial proof of concept (PoC) to validate technical feasibility, Conclusion developed a minimum viable product (MVP) in just seven weeks. This first working version allowed users to test the solution in practice. Kelly explains: "During this phase, we maintained a dedicated chat channel with our core users, allowing us to get immediate feedback: 'Take a look at this, what do you think?' That enabled us to make adjustments very quickly."

The close collaboration paid off. The entire project took just five months from start to finish, after which the fully tested solution was deployed. The 1 January 2026 deadline was met and user adoption proved remarkably smooth.

 

Nick says: "Once the MVP was ready, progress accelerated rapidly. Because we had invested heavily in usability from the start, adoption happened very quickly. The fact that end users had been actively involved throughout the process made a real difference."

 

According to Adam, the partnership between RGF Staffing and Conclusion also helped maintain momentum. "We worked with short lines of communication and were always direct and transparent with each other. Even when conversations were difficult, issues were addressed immediately. That kept the project moving and continuously improved the quality of the solution."

The introduction of the Collective Agreement Checker also became the starting point for wider innovation within RGF Staffing. According to Adam, the project delivered not only a practical solution but also uncovered new opportunities.

 

"This was one of our first genuine back-office AI projects and it set a great deal in motion. We already had ideas for improvements, but had not previously been able to realise them. Thanks to this new AI solution, we now have a technical foundation for follow-on initiatives." The project also revealed unexpected benefits. "Our business control department can now use the data generated by the tool for cost price calculations. As a result, the project ultimately delivered even more value than we initially anticipated."