With vision processing, augmented reality and the internet of things I think we are primed to live in a world where devices will be personified in AR. Personified devices will then be able to tell me what they are capable of and I will be able to query them to learn about them. I will also be able to find out what's wrong with them. I think chatbots will be the ideal interface to talk to IoT devices for most things however AR will be needed for a richer experience/in depth device instruction.

Envision a world where a google glass like device take see what you are seeing, identify what it you are looking at, figure out if it is an IoT device, notify you in AR if it is (via icon hovering over device or something), allow you to "activate" the device's personified self and talk to it, question it, assign a task to it or a variety of other interactions supported by the IoT device.