A SUITE OF ADVANCED AI TOOLS FOR CUSTOMER SUPPORT USING RETRIEVAL-AUGMENTED GENERATION (RAG)

This paper introduces a suite of customer facing advanced artificial intelligence tools (AI), thoroughly designed and developed on the innovative Retrieval-Augmented Generation (RAG) architecture. The system delivers a “digital apron in the aisle” experience—clear, step-by-step guidance grounded in verified company knowledge (product catalogs, installation guides, safety content, and store policies). We describe the ingestion and retrieval pipeline, LLM prompting with guardrails, and an evaluation framework tied to support KPIs. Four detailed UC1-style DIY scenarios (paint, tile, thermostat, and deck repair) demonstrate how the assistant improves confidence, accuracy, and outcomes for home-improvement customers. We also added a layer of abuse/sensitive-text filter, a competitor item or product filter restriction model, and a STRIDE-based security threat model.