PubNub Insights has introduced a new "Group by Channel Patterns" feature, enabling users to seamlessly transition between macro trends and detailed metrics by grouping individual metrics based on channel patterns. This allows for optimized platform performance and informed decision-making. The feature was developed considering diverse customer industries with unique channel naming conventions, varying channel volumes, and the need for real-time performance. To identify meaningful patterns in channel names, PubNub explored several methods including Suffix Trees, Trie-Based Approaches, and Naïve Log/Template Discovery, ultimately choosing a combination of Tokenization and TF-IDF for its balance of performance, cost, and immediate availability. Artificial Intelligence, though considered, was deemed unsuitable due to concerns over customer data privacy, excessive processing requirements, and cost. The chosen approach efficiently tokenizes channel names to surface significant patterns, providing a lightweight, transparent, and fast solution without requiring AI integration. The flexibility of this feature accommodates the varied naming conventions of different customers, allowing for custom pattern additions. The "Group by Channel Patterns" feature is now available for Insights users, offering an enhanced data exploration experience.