Question-answering and MRC conversational agents

Bots can be trained to answer questions based on information contained in a knowledge base (KB). This setting is called the question-answering setting. Another related area is machine reading comprehension or MRC. In MRC, questions need to be answered with respect to a set of passages or documents provided with the query. Both of these areas are seeing a lot of startup activity and innovation. A very large number of business use cases can be enabled with both of these types of conversational agents. Passing the financial report to a bot and answering questions such as the increase in revenue given the financial report would be an example of MRC. Organizations have large digital caches of information, with new information pouring in every day. Building such agents empowers knowledge workers to process and parse large amounts of information quickly. Startups like Pryon are delivering conversational AI agents that merge, ingest, and adapt...