Classifying the intent of a text is one of the most essential goals of Natural Language Processing. It can be applied e.g. in customer service centers, to quickly route messages to the correct teams that can answer questions in short time. Alternatively, it is heavily used in chatbots that have pre-defined answers, but need to match the answer to the question.
Combined with a Named Entity Recognition, such models can be leveraged to automatically answer personalized questions. E.g. if a client asks about the delivery status providing their reference number, a simple connection to the respective management system can provide answers without a human in the loop.
Building such systems requires large-scale datasets, so that your model achieves precise prediction quality. refinery helps you to scale and manage such datasets efficiently, in hours instead of weeks.
If you’re interested in trying refinery, install our open-source version. We’re more than happy to help you along the way if you have any questions.