Services
Industries
Software Engineering
Cyber Security
IT Consulting
Artificial Intelligence
Progress OpenEdge
Cloud
The global debt collection services market is projected to exceed 41 billion USD by the early 2030s, growing at a steady 2.7 to 3 percent annually. As a result, organisations increasingly rely on smarter, data-driven tools to stay competitive. This means analysing existing debt portfolios to recover them more effectively and communicating with debtors in the most efficient way possible.
Legal Balance recognised this opportunity and chose Baltic Amadeus to further strengthen their analytical capabilities.
Legal Balance is a debt management and recovery service provider. They also buy and administer debts from small and medium-sized enterprises, large companies, and various organisations.
In this field, it is important to answer several core questions: How to identify debts most likely to be repaid? How to group debts by repayment likelihood, recovery time, and other factors? Legal Balance had internal processes to address these questions, yet they saw an opportunity to strengthen them. Their aim was to make debt recovery and communication with debtors more accurate, more efficient and data-driven, and more consistent across the organisation. That’s when we stepped in as their partner.
To start the process, the Baltic Amadeus team brought together cross-functional expertise in data science and AI engineering and began by analysing all available debtor and payment data. As a result, our team created a unified dataset with around 70 meaningful features per debtor, forming a solid foundation for accurate machine learning modelling.
The Baltic Amadeus team then trained and refined the machine learning models through several iterations, testing different approaches and focusing on what delivered the most consistent results. This led to an ML-based model that assigns debtors to repayment likelihood segments.
Building on these predictions, an LLM-powered assistant was developed to support everyday recovery decisions. The assistant brings together all relevant case information, applies Legal Balance’s business rules, and helps managers quickly understand each situation, decide on next steps, and communicate more effectively. The solution runs as a Python-based service with an internal interface and is designed for future integration into the organisation’s CRM.
The AI project helped Legal Balance gain a clearer picture of repayment risk and recovery potential, both for individual cases and for entire debt portfolios. By bringing different data sources together, the client could make confident data-driven decisions when planning recovery strategies and prioritising their efforts.
The prediction models supported estimates around repayment likelihood and timing, while the AI-powered assistant helped with day-to-day recovery work. Managers could review cases faster, focus their efforts where it mattered most, and choose recovery and communication approaches more consistently.
As a result, Legal Balance improved how decisions were made across their recovery operations. The solution reduced manual analysis, made daily work more efficient, and created a strong foundation for future automation and CRM integration.
