Kaspar Solutions

The client

Kaspar Solutions was founded in 2008 in Heist-op-den-Berg. They build software for case and file management in the legal sector, with a strong focus on judicial officers. Their platforms bring the parties involved together, from courts to clients and partner organisations, through webservices and web interfaces. In that context, security is non-negotiable.

Under the motto "Smart, Secure Services", Kaspar Solutions offers clients a complete ICT service: from analysis, advice and design through to development, implementation, support and training.

The case

AI is changing the way software is built, and Kaspar Solutions has already been experimenting with it extensively. However, this was previously done on an ad hoc basis: everyone in their own way, with results that were difficult to compare and knowledge that was hard to share. The question was therefore no longer whether AI would have a place within the team, but how to approach it in a structural and controlled manner.

This challenge applied to both analysis & product ownership and development. On the analysis side, there was a need for more structure and version control in documentation so that it could become a reliable knowledge base for AI. At the same time, Kaspar Solutions wanted to support colleagues in a uniform way of working with AI. Within development, AI had to become part of the daily development process itself, not a side experiment.

The goal is clear: to build a way of working together that Kaspar Solutions can continue to use and refine in the long term. Euricom supports them in this with expertise, hands-on guidance, and knowledge transfer.

Our approach

Together with Kaspar Solutions, Euricom set up an AI-supported software development lifecycle around Claude Code. Two Euricom consultants are working within the team for this purpose, each with their own focus.

Analysis and product ownership. Together with the team, Kevin is restructuring and versioning documentation to make it effectively usable as a knowledge base for AI, without disrupting existing workflows. In addition, he mentors and trains colleagues in a uniform, controlled approach to AI usage and translates functional analyses into custom AI solutions.

Development. Matthias introduces AI into the development process itself, in close consultation with the relevant team leads. The technical foundation for this: .NET Core, SQL, Angular, and Azure DevOps.

Training and opt-in adoption. Euricom designed and delivered two training days, tailored to each role: 27 August 2026 for developers and 28 August for analysts and other roles. The focus was on controlled, responsible use of AI: working human-in-the-loop, and the principle that a person stays responsible for every result. Afterwards, everyone could request a Claude seat on an opt-in basis. Kaspar Solutions wanted to give its people the chance to work with AI, without imposing it.

In AI adoption projects, Euricom always focuses on knowledge transfer. What the team continues to do independently after the project is more important than what happens with AI during the project itself.

The result

Within three weeks of the training days, Claude is part of daily production work at Kaspar Solutions.

• 41 Claude seats, 29 of them active every week.
• 162,893 lines of code generated with Claude Code and accepted by developers in the first 18 days of September 2026, across a development team of 18.
• 205 Claude Cowork sessions in the last 30 days, up 469%. Daily Cowork users rose from 0–1 in June and 3–4 over the summer to peaks of 8–10 in September.
• 56% of monthly Claude Code and Cowork users use them every day.

Usage accelerated from early September, about a week after the training days: the time between requesting a seat and getting to work with it.

The impact now reaches beyond development. The PMO has asked for Claude artifacts to track analyses and for skills that generate reports, which is already saving time on reporting. Next up: connecting Cowork to Azure DevOps, so project data no longer has to be exported as CSV files.

"We wanted to roll out AI in a controlled way, bring everyone on board and make sure everyone knows the etiquette. We tailor training to each role and relevant to everyone's daily work, because AI only pays off when our people use it with confidence and responsibility."

- Bernd Vertommen, Lead Architect, Kaspar Solutions