In an interview with Portfolio, Sebastian Clajus, Behavioral Director at PAIR Finance, explained how the company uses a mix of artificial intelligence and behavioral science to handle overdue debts. Clajus argued that the common dichotomy — that a debtor either cannot pay or refuses to pay — is overly simplistic; shame, stress, decision overload and poorly designed processes often play a major role.
Why behavioral science matters
According to Clajus, many debtors are not willfully noncompliant but are constrained by what economists call scarcity mindset: shame, stress and limited cognitive capacity can prevent productive action. PAIR Finance focuses on changing processes rather than people — reducing friction, removing obstacles and simplifying communication so settling a debt becomes less stressful and more feasible.
When AI is appropriate and when human empathy is needed
PAIR Finance uses AI widely but follows a clear division of labor between machines and people. Advantages of AI highlighted by Clajus include:
- lowering perceived stigma, since clients may feel less judged when interacting with an automated system;
- handling routine tasks effectively (requests for installment plans, payment deferrals, payment confirmations).
The company currently resolves roughly 70 percent of cases through technology and AI without human contact. However, human involvement remains essential when a genuinely empathetic response is required — for example, when a client is emotionally distressed, wants to complain, or needs a more personalized, human connection.
What data is used to personalize journeys?
Personalized customer journeys are built on several layers: basic information (country, language, cultural context), channel preferences (channel effectiveness can vary by market — e.g., SMS may work better in Germany while WhatsApp could be stronger in Austria), and behavioral data collected during the process (email opens, visits to the payment page, previous interactions).
PAIR Finance operates in 13 countries and runs many experiments to understand local differences. A simple example: if a client successfully resolved a previous claim, the system can identify that pattern and send messaging that leverages that prior success — often only a few well-chosen data points are enough to create a more personal and effective path.
Technical and ethical safeguards
The platform uses in-house machine learning and AI technology. It learns patterns from about 10 billion data points drawn from a GDPR-compliant framework with ISO-27001 certification. PAIR Finance emphasizes a “human in the loop” approach: the most important decisions are made by people, textual content is prepared by the behavioral team, and the AI suggests intents, patterns, channels and timing — but does not independently generate content.
Measuring effectiveness
Effectiveness is tracked through classic business KPIs and customer experience metrics. Recovery rates for corporate clients are on average 10–15 percent higher than traditional providers. The company’s Trustpilot rating is 4.7. Automation rate (about 70 percent of cases handled without human intervention) is another key metric for efficiency and scalability.
European adoption and the differentiator
Clajus notes that while many organizations claim to use AI, the term covers a wide range of practices — from casual use of ChatGPT to fully AI-native platforms. PAIR Finance started as an AI-first company and combines that with explicit use of behavioral science. He observes that behavioral approaches are more established in some markets (Scandinavia, France, the UK) than in others (Germany shows relatively lower explicit use in collections).
PAIR Finance’s view is that AI delivers scalability and speed, while behavioral science supplies the psychological insight that makes solutions effective and humane.
Presentation and conference
Sebastian Clajus will speak about these topics in greater detail at the Követeléskezelési Trendek 2026 conference, organized by Portfolio and EOS, on September 15. Organizers are already accepting registrations.
Tags: artificial intelligence, behavioral science, debt collection, automation, customer experience, GDPR, ISO-27001
