Imagine having instant access to a vast library of knowledge while helping customers. That's the power Comcast is giving its agents with "Ask Me Anything" (AMA), an AI tool that leverages the magic of large language models (LLMs). Customer service is a balancing act. Customers crave personalized support, but providing that at scale gets expensive. While chatbots can handle simple tasks, many situations still demand the human touch, especially for complex issues or sensitive topics. This puts immense pressure on agents to quickly find the right information while juggling multiple conversations. Comcast's AMA aims to ease that burden. Instead of frantically searching through internal documents, agents can simply "ask" the LLM a question and get real-time answers, reducing the need to switch between different applications. This retrieval-augmented generation (RAG) system combines internal knowledge sources with the power of LLMs. It works by efficiently indexing knowledge articles, retrieving relevant chunks of text based on agent queries, and then using a "reader" LLM to generate concise answers with citations. Comcast's approach goes beyond standard RAG implementations. They've fine-tuned the system with synthetic data to improve search relevance and ensure agents get the most accurate information. A key feature is the "Citation Rail," which provides agents with direct links to the source documents, boosting confidence in the AI's responses. Early trials of AMA have been impressive. Agents using the tool saw a 10% reduction in handling time compared to traditional search methods, translating to potential millions in annual savings. Agent feedback has been overwhelmingly positive, with an 80% approval rating. The system is now being rolled out to thousands of agents, transforming how Comcast delivers customer support. The journey doesn't stop here. Comcast is continually refining AMA, including A/B testing a new reranking module to further enhance search accuracy and reduce instances where the system can't find an answer. This is a prime example of how AI can empower human workers, not replace them. By giving agents the information they need at their fingertips, Comcast is creating a better experience for both its employees and its customers. As LLMs continue to evolve, expect even more innovative applications like AMA to emerge, reshaping the future of customer service and beyond.