novobanco GenAI Customer Complaints Hub Faster answers. Smarter service. Stronger trust. Qorus-Infosys Finacle Banking Innovation Awards 2025

Concept and Objectives The novobanco GenAI Customer Complaints Hub is a proprietary AI framework designed to automate and enhance the full lifecycle of customer complaints — from handwritten intake to resolution. It combines Large Language Models (LLMs), Retrieval-Augmented Generation (RAG), OCR Intelligence, and agentic AI to deliver fast, accurate, and compliant responses. Reasons Behind In the banking sector, complaints are emotionally charged moments that can either erode or restore trust. novobanco receives a high volume of complaints, many of which are still handwritten at branches. These are often difficult to interpret and require manual transcription by branch Relationship Managers. The GenAI Hub introduces OCR Intelligence to interpret these documents with high accuracy, enabling immediate resolution or intelligent routing to the most competent human agent, bringing relevant data to context for thorough adequate responses. Future widening of available data sources will allow for an increasingly holistic view of the customer and subject matter expertise critical to a complete and robust handling of responses. State of Competition While many banks rely on generic chatbot platforms, novobanco has developed a fully in-house, domain-specialized solution. This ensures full control over data, compliance, and quality — and allows for deep customization to the bank’s operational reality. Sources of Inspiration The project was born from the limitations of existing tools and the ambition to lead in AI-driven transformation. A successful proof of concept in early 2024 demonstrated the potential of GenAI to reduce resolution times by up to 80%, validating the decision to scale the solution. Departments Involved The project was co-led by the Transformation and Innovation Office, Data and Customer experience areas with direct involvement from: - Information Technology and Data Science - Legal, Risk, and Compliance - Customer Service Operations - Data and Analytics (for OCR and handwriting interpretation) Main Results So Far - Completion of a PoC with up to 80% reduction in processing time for simple complaints. - Development of a scalable architecture using Azure and OpenShift. - Implementation of a handwriting interpretation module tested on real complaints. - Launch of the first internal AI assistants for complaints answering through an application powered by AI agents.