NSIS Implements Internal AI: Inquiry Processing Speed Increases 20-Fold, Staff Unchanged

NSIS (AO 'National Insurance Information System') has implemented an internal AI to modernize business processes and transform its automated insurance information system into an intelligent platform. The AI operates within the company's secure perimeter due to data security requirements. In the first phase, inquiry handling was optimized, resulting in a 20-fold increase in processing speed. Despite an average monthly growth in inquiries of 11% and an annual doubling, headcount remained unchanged because of the AI-powered processing system. A large language model deployed on local NSIS hardware serves as a key tool, enabling RAG (Retrieval-Augmented Generation) for technical support, with future adaptation for other departments. Developers can now upload custom data to local LLM models, creating corporate scenarios. For testers, an agent and MCP server have been deployed to create an integrated development and test management environment, accelerating release cycles. The company's strategy emphasizes 'AI as an assistant, human as captain,' with AI handling routine analysis while architects make decisions. Future plans include AI-assisted code review for security and quality issues. The transformation aims to free employees from routine tasks, not replace them, and has already evolved the Automated Insurance Information System (AIIS) into a full analytics platform supporting reinsurance, KBM calculations, and fraud detection.