A method and system are provided. The method includes automatically monitoring one or more voice communication sessions occurring on multiple communication devices to transcribe into text, by a processor-based voice-to-text transcriber, voices uttered by two or more respective participants other than a user in each of the voice communications sessions. The method further includes comparing, for each of the voice communication sessions, the topic of interest to conversation features derived from the text to detect a match there between. The method also includes providing the user with a summary of each of the voice communication sessions for which the match has been detected by said comparing step, the summary being based on the text. The method additionally includes automatically connecting the user to at least one of the voice communication sessions for which the summary thereof has been provided to the user.