Virtual Assistants are adding dialogue systems that add a smart language component. They work with natural language components that enables systems to provide responses that are more conversational. Automatic speech recognition, video recognition, gesture, gazing, graph, and interaction models can be added that captures information and runs through the systems to add more to the personable experience. Facebook is using some of these capture methods to build smart systems that learn from instruction by teachers (Bohouta and Këpuska, 2018)
Amazon rolled out a contest some time ago challenging groups of students to build better conversational bots. The teams deployed various natural language, speech recognition, and dialogue models to meet the challenge. The technologies created helped Alexa broaden scope to millions of customers. The socialbots injected conversational intent tracking, inappropriate and sensitive content detection, conversational quality evaluation, and even socialbot scalability (Chandra, Anu, Behnam, Ashwin, Raefer, 2018).
DVA's have the advantage to search data from many cloud systems, internet, databases, knowledgebases, and custom processes and instruction. In many cases, asking the DVA to look for something could be faster than looking up using a specific apps. A good example is some assistants, especially Google Assistant, can interact with Twitter. Asking the assistant to look for conversations or posts can be quicker than going to Twitter and scrolling or applying manual searches.