The customer would like dedicated AI Search Analytics to provide deeper visibility into how users interact with LLMs such as Claude and ChatGPT, and to measure the effectiveness of AI-assisted search. They are looking for insights such as whether the AI successfully answered a query, queries that resulted in poor or no answers, overall usage trends, recurring search patterns, and content gaps where the AI was unable to provide accurate responses. These analytics would help their team continuously improve documentation, evaluate AI answer quality, and measure AI Search adoption.

Additionally, they would like analytics that differentiate traffic originating from LLMs versus the traditional Help Center search experience. For example, a high-level breakdown showing that 60% of searches originated from LLMs (e.g., Claude, ChatGPT) and 40% from the Help Center, along with usage trends over time. This would help them understand how users are consuming their documentation across different channels and assess the ROI of AI-driven knowledge discovery.