Work

Developing AI‐driven decision support for the proactive identification and evaluation of financially distressed companies by Belgian commercial courts

Year: 2026

Type: article

Abstract: The early identification and accurate evaluation of financially distressed companies by commercial courts is crucial for maintaining a healthy economic fabric. This enables timely support for struggling companies with recovery potential while also facilitating the swift winding up of those unlikely to recover. Despite its economic importance, the use of artificial intelligence (AI) in this context remains limited. We investigate how an AI‐based decision support system can be designed and validated to support commercial court judges in identifying and evaluating financially distressed companies. Specifically, we train predictive machine learning models that combine red flag data (e.g., social security debt, debt under seizure) with annual account data to predict whether a company will be referred for bankruptcy proceedings and to rank companies by their likelihood of referral to enable prioritization. Results indicate moderately strong performance, underscoring the potential of AI‐assisted decision‐making. Importantly, the proposed system does not replace judges, but supports them in allocating scarce resources more effectively.

Source: International Transactions in Operational Research

DOI: 10.1111/itor.70267

Authors: Stijn Van Ruymbeke, Joke Baeck, Klaas Mulier, Thomas Demeester

Institutions: HOGENT University of Applied Sciences and Arts, Ghent University, KU Leuven, iMinds

Language: English

Topic: Financial Distress and Bankruptcy Prediction

Subfield: Accounting

Field: Business, Management and Accounting

Domain: Social Sciences

Open Access status: hybrid

Estimated APC paid: $4,030

Funder: Vlaamse regering