Every year 2nd line has to manually review the risk assessments staged by 1st line. Because this is very time intensive, our sponsor contacted us with the question “how can we do better?”. To find a solution we held a short series of experiments with NLP. In these experiments we had to overcome the challenges of subjective language by different raters and their difference in the usage of words. Eventually we created OpBot: our AI solution that is able to prioritize the risks on the chance to be anomalous. In this way 2nd line can set priorities in their risk evaluation.


Given OpBot is still in the prototype phase: – 2nd considered this first experiment with AI as success and want to identify new use-cases for AI. – Over 50% of the risk assessments that OpBot indicated as anomalous was considered as anomalous bij 2nd line.

Curious what we can do for your organisation?

Would you like to know more this case or what the Amsterdam Data Collective can do for you? Get in touch with Casper Rutjes at casper@amsterdamdatacollective.com, or check our contactpage.