With the combination of highly accurate speech to text, intelligent keyword/topic extraction, and predictive insights training models, automatic event detection in spoken content is not only possible, but works really well.