The increasing complexity of industrial robots, driven by the advancements of Industry 4.0, requires efficient troubleshooting solutions to address numerous hardware and software malfunctions.

This research proposes a domain specific, AI-powered chatbot leveraging Retrieval Augmented Generation (RAG) to streamline technical support within ABB’s Robotics Division. The chatbot unifies fragmented data from Salesforce, Robdesk, and Azure DevOps, employing advanced Natural Language Processing (NLP) techniques, such as semantic search and vector embeddings, to improve the accessibility and relevance of technical information.

The study focuses on developing a robust retrieval pipeline using Azure Cognitive Search and OpenAI models, supported by comprehensive data cleaning, conversation summarization, and metadata normalization. Quantitative evaluations reveal a 40–50% reduction in the average time to retrieval compared to manual searches, significantly improving efficiency. Qualitative feedback from robotics engineers highlights the improved clarity, relevance, and usability, validating the ability of a chatbot to provide consolidated, context-aware troubleshooting solutions.

Key contributions include optimizing data preparation workflows, integrating heterogeneous repositories, and addressing robotics specific terminology. The scalable system architecture allows for future expansion to additional data sources and realtime diagnostics. This research demonstrates the transformative potential of RAGbased systems in industrial robotics, offering a foundation for broader applications in other technical domains.