AI Transformation in Regulated Payment Services
Jakob von Ingelheim from UNSA details how a European payment provider leveraged regulatory process catalogs to identify high-impact AI use cases. The strategy reduced onboarding time by 90% and improved fraud screening efficiency. Key insights include the shift from human-in-the-loop to human-on-the-loop and the strategic preparation for agentic commerce.
Strategic Integration of AI in Regulated Environments
UNSA, a European payment service provider with over 90,000 merchants, has successfully navigated a complex AI transformation by leveraging its regulatory obligations as a strategic asset. Rather than viewing compliance as a burden, the company utilized its mandatory process catalogs to map out high-impact automation opportunities. This disciplined approach allowed UNSA to move beyond superficial AI applications and embed intelligence directly into core operational workflows, resulting in measurable efficiency gains and risk reduction.
Operational Efficiency and Process Optimization
The most significant operational breakthrough occurred in merchant onboarding. By decomposing the Know Your Customer (KYC) process into granular steps, UNSA applied AI to automate data extraction and verification. This reduced the average onboarding time from 47 days to just 8 days, a 90 percent improvement. Similarly, in fraud prevention, AI-driven adverse media screening reduced manual review time from 15 minutes per case to under three minutes, without compromising detection accuracy. These improvements directly translate to faster merchant activation and reduced operational costs.
Customer Experience and Multilingual Support
UNSA has also transformed customer support by deploying AI agents capable of handling 50 percent of inquiries autonomously. A notable use case involves a WhatsApp-based AI agent that offers installment plans to customers in payment difficulties. By supporting 37 languages, this solution removes linguistic barriers and reduces the psychological friction of discussing debt, leading to a 30-35 percent increase in successful resolution rates. This demonstrates how AI can enhance customer experience while simultaneously improving financial outcomes.
Cultural Adoption and Leadership Accountability
A critical component of UNSA’s success is its focus on cultural adoption. The company implemented an "AI Week" to reduce employee anxiety and a hackathon that resulted in over 100 internal apps and 60 agents built by staff. Crucially, UNSA mandated that all senior leaders build their own AI agent as a performance goal. This hands-on requirement ensures that leadership possesses a practical understanding of AI capabilities, fostering a culture of experimentation and informed decision-making.
Future-Proofing for Agentic Commerce
Looking ahead, UNSA is positioning itself for the rise of agentic commerce, where AI agents will execute purchases on behalf of consumers. The company is developing secure, auditable payment infrastructure to support this shift, ensuring that transactions remain compliant and trustworthy. By proactively addressing this emerging trend, UNSA aims to maintain its competitive edge in a rapidly evolving digital landscape.
Key insights
-
Regulatory process catalogs can serve as a robust framework for identifying AI automation opportunities. This ensures that AI initiatives are aligned with compliance requirements and risk management strategies.
Impact: Reduces the risk of AI implementation failures and ensures that automation efforts are focused on high-value, compliant processes.
-
Decomposing complex processes into granular steps allows for targeted AI application, leading to significant efficiency gains. This approach was key to reducing onboarding time by 90 percent.
Impact: Accelerates time-to-market for new customers and reduces operational costs, directly impacting revenue and profitability.
-
The shift from human-in-the-loop to human-on-the-loop is essential for scaling AI in regulated environments. This model allows AI to handle routine cases autonomously while humans focus on exceptions.
Impact: Improves operational efficiency and allows skilled employees to focus on complex, high-value tasks, enhancing overall service quality.
-
AI-driven multilingual support can significantly improve customer experience and resolution rates. By removing language barriers, companies can better serve diverse customer bases.
Impact: Increases customer satisfaction and retention, while also improving financial outcomes through higher resolution rates in debt management.
-
Mandating hands-on AI experience for leadership is crucial for successful AI transformation. This ensures that decision-makers have a practical understanding of AI capabilities and limitations.
Impact: Fosters a culture of innovation and informed decision-making, leading to more effective AI strategy and implementation.
Action items
-
Audit existing compliance and process catalogs to identify high-impact AI automation opportunities. Focus on processes with high manual effort and clear regulatory requirements.
Impact: Provides a structured roadmap for AI implementation, ensuring alignment with compliance and risk management goals.
-
Decompose core operational processes into granular steps to identify specific AI use cases. Prioritize steps with high data extraction or verification requirements.
Impact: Enables targeted AI application, leading to significant efficiency gains and reduced operational costs.
-
Implement a human-on-the-loop model for low-risk, high-volume tasks. Define clear criteria for when human intervention is required.
Impact: Scales AI capabilities while maintaining compliance and quality, allowing employees to focus on complex exceptions.
-
Deploy AI-driven multilingual support agents to handle routine customer inquiries and complex issues like debt resolution. Ensure the AI is trained on diverse language models.
Impact: Improves customer experience and resolution rates, while reducing support costs and enhancing brand loyalty.
-
Mandate that all senior leaders build a functional AI agent as a performance goal. Provide resources and training to support this initiative.
Impact: Builds leadership proficiency in AI, fostering a culture of innovation and informed strategic decision-making.
Quotes
“Wir sind ja reguliert. Wir haben in drei verschiedenen Ländern eine nicht ganz Vollbanklizenz, aber sozusagen man kann sich vorstellen wie eine kleine Banklizenz, eine sogenannte ZAG oder Payment Provider Lizenz.”
“In den letzten 18 Monaten fast 90 Prozent höhere Geschwindigkeit erzielen können. Das heißt, wir haben vor 18 Monaten 46, 47 Tage gebraucht im Durchschnitt für so ein Onboarding.”
“Ich habe zum Beispiel in meinem Bereich von allen Führungskräften Anfang des Jahres in die Ziele geschrieben, dass sie selber einen Agent bauen müssen, egal was es ist.”