‘The legal department of the Municipality of Amsterdam is facing increasing workloads. In recent years, the number of objections has risen sharply and case files are becoming increasingly complex. The caseload is simply too large. As a result, residents and businesses sometimes wait a long time for a decision,’ says van der Meer, Lead Data Scientist at the municipality’s Legal Office (Juridisch Bureau). ‘I wondered: how can AI support lawyers’ workflow without them losing control over the substantive handling of cases?’

Van der Meer tackled this challenge as part of his in-company graduation project for the MBA in AI, Data & Analytics. The AI tool LegalCheck helps the municipality’s legal professionals draft responses to notices of objection. A notice of objection and other relevant information are the input. The system then automatically searches a database of anonymised cases previously handled by the municipality. Within 20 seconds, the legal professional receives a draft letter for review and approval.

From pilot to workflow

‘LegalCheck has made a real impact in a short period of time,’ says Van der Meer about his graduation project, which he completed with an A+ in 2025. ‘Even during my graduation project, the system was tested and evaluated in practice as a pilot. At that time, the system worked for 3 types of objections and 17 legal professionals used it. Afterwards, the procedure for 7 types of objections were added. We can see that LegalCheck works well across all cases in different legal domains.’

The pilot lays the foundation for a cloud platform that makes the application an integrated part of the daily workflow of the municipality’s 150 legal professionals. The system is also ready to be deployed for all types of objection procedures handled by the municipality, around 100 in total.

More than efficiency gains

The pilot evaluation shows that LegalCheck delivers significant efficiency gains. ‘The tool can reduce the time needed to prepare legal advice by 50 to 70%. Legal professionals are also enthusiastic about the quality of the draft letters. In over 99%of cases, all relevant facts are already correct. The generated output is directly usable for roughly 80 to 95%of the content. Legal professionals regularly say: I could hardly have written this letter any better myself.’

According to van der Meer, another important outcome is that LegalCheck improves the consistency and quality of legal decision-making. ‘The system searches a database containing thousands of cases. No person can oversee that amount of data. As a result, people may sometimes give different advice in similar cases. LegalCheck’s letters are more consistent and can reduce bias. In addition, the letters provide a more detailed explanation of the grounds for objection. And the tone is more empathetic.’

‘The added value for the organisation goes far beyond efficiency gains. LegalCheck can make a real difference to municipal services: residents and entrepreneurs receive clarity about the handling of a notice of objections sooner,’ van der Meer emphasises. ‘The pilot project also starts an internal learning process. For example, we eventually want to analyse whether issues already arise earlier in the decision-making process. This could help prevent objection procedures altogether.’

Bridging technology, business and societal impact

Julien Rossi, lecturer in Language Technology for the MBA in AI, Data & Analytics, was closely involved in the development and implementation of LegalCheck as thesis supervisor. ‘As with other graduation projects, it is clear that the technical side of AI isn’t the biggest challenge. The point is to convince all stakeholders that your AI solution works well and to monitor the effects carefully. Attention is therefore needed for the organisation’s entire ecosystem. Who are the stakeholders? What resistance exists? What are the requirements set by laws and regulations?’

‘The MBA in AI, Data & Analytics builds a bridge between technology, business and societal impact,’ adds van der Meer. ‘You take courses on the technical and mathematical side of AI, such as programming and machine learning. But also on business strategy, leadership, governance and compliance. What is technically possible with AI models and what isn’t? What does a successful development process look like? Which ethical dilemmas arise? You learn to approach challenges from different perspectives. That allows you to use the power of AI in a safe and responsible way.’

A technical solution that adds value

Many organisations are experimenting with AI, but find it difficult to scale applications and embed them structurally into day-to-day work. Students in the 2-year part-time MBA in AI, Data & Analytics work on exactly these kinds of issues. The goal: translating academic knowledge into concrete tools that deliver societal value.

According to Rossi, one of the biggest challenges for the LegalCheck initiative was carefully addressing legal requirements and societal values. ‘Users quickly recognised that the tool simplifies and accelerates the Legal Office’s work. However, the municipality naturally sets very high standards for compliance with laws and regulations. In addition, the rights and interests of citizens come first. That’s why compliance by design and privacy by design are central to the project. LegalCheck is therefore much more than a technical success. The 360-degree perspective of all stakeholders was crucial.’

‘LegalCheck works well technically and genuinely adds value to the daily practice of the municipality’s legal professionals,’ says van der Meer about the project’s success factors. ‘The AI application was developed bottom-up. Users understand how LegalCheck works; it is not a complete black box. The tool is aligned with daily workflow, so legal professionals experience it as a valuable support for their work.’

AI and people

‘The LegalCheck pilot shows that AI can be deployed responsibly within the public sector,’ says van der Meer about the practical experience. ‘The system can be further developed with new functionalities. We’re also exploring expansion to other municipal departments, and eventually to other municipalities and public organisations. There is enormous potential for a solution in which AI and people strengthen each other.’