Finanzmarktaufsicht Liechtenstein

With “Fit & Proper”, the FMA uses bots and AI to screen prominent individuals for critical reporting based on public data. First, the bot carries out the website search and downloads the pages with the search hits. Next, the AI checks the hits for controversial content and ensures that they refer to the person in question. The results are presented to the
experts in a user interface to verify the hits marked as controversial and to continuously improve the AI through feedback.

Fit & Proper was trained with a
dedicated data pool

It was not possible to draw on existing sentiment analysis models for the implementation, because these models are only able to evaluate detected content negatively or positively. "Fit & Proper", on the other hand, is only intended to highlight controversial content – the employees themselves carry out the negative or positive assessment. Initially, a data pool of 200 individuals was established: Each person was searched for on the web and all hits were evaluated manually according to two criteria: Is it the right person? And is the content to be considered controversial? By evaluating the sample, the program learned the criteria for identifying which words or word combinations are classified as controversial, and created a rating grid.

From workshop to solution in four months

In a joint workshop with the FMA, ti&m’s AI team identified potential areas of application for artificial intelligence. Based on the defined use case “Fit & Proper”, we successively developed a proof of concept and also outlined possible further fields of application for AI. In four months, the proof of concept was developed into an integrated solution. This runs as a microservice application on Kubernetes infrastructure at the customer’s site. The solution was initially available as a web app and was integrated into the FMA’s core system in a second step. Since the initial implementation, the FMA and ti&m have collaborated closely to develop “Fit & Proper” on an ongoing basis. In doing so, they continuously evaluate and implement the potential of new technological developments to optimize the existing solution.

Unbiased and faster results

When appointing key positions, such as members of the Executive Board, the Board of Directors or compliance officers, the FMA has the task of screening the person concerned and approving the mandate. Screening of individuals is a resource-intensive process in which information from various sources must be collected, evaluated, and filed. The “Fit & Proper” check on public search engines provides relevant information about individuals from the web.
This supplements the information collected from criminal and debt enforcement register extracts and professionally managed databases. The digitalization of this process has standardized and simplified collaboration across all departments. The automated process has eliminated the previously laborious process steps caused by media discontinuity. All results are now automatically saved and archived as a PDF. The processing of the verifications can thus be completed 30–50 percent faster. In addition to the time saved, the quality of the results has also improved. For one thing, “Fit & Proper” carries out the web search without the bias that might occur when employees use their browsers. Secondly, the tool automatically processes search hits in any language and shows the user a translation of the website in German and English.

Want to learn more?

Read our blog post on efficient background screening using AI!

To the blog post

“Screenings can now be completed much quicker, thanks to the fully digital and partially automated process. In addition to the time saved, the quality of the results has also improved.”
– Roger Guntli, Head of Information and Communication Technologies for the FMA

Head of Products & Banking Innovations

Fabian Braunwalder

Are you interested in the “Fit & Proper” check on public
search engines?

Fit & Proper was trained with a dedicated data pool

From workshop to solution in four months

Unbiased and faster results

Want to learn more?

Fit & Proper was trained with a
dedicated data pool