Hello. For discussion, I am wondering about how dynamic data (e.g., the weather), responsive design, alerts, and notifications can factor into chatbot architectures.
End-users might ask a chatbot about a weather forecast for one or more regions of interest to them. The chatbot might provide the users with answers but, shortly thereafter, a weather forecast could change, the backing data could change.
How might we design architectures so that AI systems could apprise the user that backing data relevant to a question that they asked has changed and/or can provide them with an updated answer? Possibilities include that: (1) a mixed-initiative chatbot could conversationally apprise the user, (2) a chatbot could make use of another communication channel (e.g., SMS, email) to apprise the user, (3) a chatbot could apprise the user using a Web browser or operating system notification system, (4) some other technical approach.
I also think about these topics with respect to moveable content. Users might clipboard or drag-and-drop content from a chatbot dialog, e.g., a weather forecast (perhaps an adaptive card), and place it into an office document. Should the replicated content, too, have some sort of updatability enabled? Might the content be able to provide real-time informational, warning, or error messages pertinent to its veracity and freshness? Or, perhaps, might it have a refresh button on it?
Thank you. Are any others here interested in chatting with AI systems about both static and fluid, dynamic, updateable, backing data?