GenAI-Enhanced Intellectual Capital Development for Effective Case Deflection and Support

Case deflection refers to the ability of customers to resolve issues independently, reducing the need to contact customer support and open cases. The success of this strategy depends on a robust knowledge management system that identifies and addresses content gaps, ensuring accessible, accurate, and comprehensive documentation for stakeholders, including customers, partners, and engineers. This paper introduces a systematic approach to developing Intellectual Capital (IC) by analyzing existing documentation and engineering case notes. A predictive model, utilizing supervised learning and natural language processing (NLP), is designed to classify bugs reported in customer support cases based on their potential for digitization into IC. By prioritizing high-potential bugs, this model allows engineers to focus on relevant cases, streamlining workflows, optimizing resource allocation, and significantly increasing IC digitization rates.

Additionally, the process incorporates large language models (LLMs) for automated IC document generation. Leveraging Retrieval-Augmented Generation (RAG), the model gathers relevant context from engineer case notes, enabling LLMs to produce consistent, high-quality, template-based documentation. This automation reduces manual effort, improves standardization, and ensures the timely creation of IC assets. By integrating AI-driven predictive modelling with LLM-based automation, this approach enhances decision-making, reduces costs, boosts customer satisfaction (CXSAT) by improving self-service resources, and accelerates mean time to final resolution (MTFR) through effective case deflection strategies. The outcome is a streamlined operational framework, greater efficiency, and a significantly improved customer experience.