Customer Case Study
In less than six weeks, Micoll and Clairiti B.V. built an H&S (Health & Safety) risk assessment agent that helps safety professionals work faster and more consistently. Time savings per analysis: around 50%. And the client started building their own agents shortly after. This is how it happened.
From domain expertise to a working RI&E agent in 6 weeks
In less than six weeks, Micoll and Clairiti developed a working RI&E agent that helps safety professionals reach a first analysis faster and with greater consistency. That pace was only possible because the project started with a thorough discovery of the existing process, including the steps, formats and working methods behind it. That foundation proved essential to translating the work into AI quickly and effectively.
At a glance
Why it worked
What Micoll contributed
- Deep subject-matter expertise and years of practical experience in RI&E.
- Real-world cases and fast validation grounded in the actual workflow.
- Intensive feedback and a rapid increase in AI understanding throughout the iterations.
What Clairiti added
- A thorough discovery of the process, steps, formats and working methods.
- Structuring the data into an AI-ready layer.
- Development of the agent in Microsoft Copilot Studio for use within M365 Copilot.
- Iterative refinement of functionality, language and UX.
- The technical translation of domain expertise into a working agent.
The full story
The collaboration between Micoll and Clairiti started with a clear ambition: not only to use Micoll's existing RI&E knowledge more intelligently, but to translate it into a concrete tool that adds immediate value to the day-to-day work of safety professionals.
A key part of the success lay in how the project started. Rather than building immediately, Clairiti began with a thorough discovery of the existing process and all the steps, formats and working methods that came with it. That made it possible to properly understand the logic behind Micoll's work and then translate it quickly into an AI solution that fit the reality of the job.
What made the project strong was the intensity of the collaboration. Instead of spending a long time designing on paper, Micoll and Clairiti worked in short iterations toward something that could be tested in practice. Domain expertise and technology were not developed one after the other, but together. That made it possible to put the agent live within six weeks.
For the client, it was striking to see how quickly the translation could be made from a highly specific RI&E and Health & Safety domain into a usable AI application for that same field. The collaboration not only strengthened the product, but also accelerated the client's understanding of AI.
The outcome was concrete. The RI&E agent reduced the time per analysis from roughly two hours to about one hour, with a first draft ready in around fifteen minutes and then time left for review and fine-tuning. At the same time, the agent helped make analyses more consistent and more complete by systematically asking follow-up questions and leaving less room for important steps to be skipped.
A second outcome was that the collaboration visibly accelerated AI knowledge on the client side. The client quickly began writing agent instructions independently and soon after that also developed alternative agents, based on the design and instructions of the agent that had been built together. So the project delivered not just an agent, but also a remarkably fast upskilling journey.
This first phase proved not only that the approach was technically feasible, but also that co-creation between domain experts and AI specialists can lead to tangible results in a short period of time.