Abstract: The present technology provides real-time topic extraction and summarization technology for communications channels and messaging platforms. The topics are extracted using a language model that extracts topics from chat log data. A topic that is selected causes the conversation panel to jump to the start of the conversation associated with the selected topic. The messages that are associated with the selected topic are highlighted or segregated from the rest of the messages. The selection of the topic and other types of engagement with topics are stored to build up a database that is used for generating a personalized ranking of topics for respective users. The quality of the topics is scored using a different language model.