I do not see any magic happening in the next decade, but we will certainly see wider commercialisation of and dependence on AI.
Even today, we depend on AI for multiple purposes. Google tends to show us what we wish to find, Quora shows us relevant questions, Translators and OCR software allow us to understand paper documents in different languages, etc.
One thing that certainly will happen is the advent of virtual assistants. There already is a lot of research in the area but the assistants in their current state are limited in their capabilities. Within a decade, I see them becoming much more capable.
In technical terms, I guess we will advance the most in the areas of Deep Learning, Natural Language Processing and Artificial Neural Networks. The first one because it introduces a layer of abstraction in Machine Learning and removes the requirement for Feature Engineering. In a sense, that makes AI more intelligent.
Natural Language Processing will continue to draw a lot of investment as it will revolutionise Human-Machine Interfaces; and finally Artificial Neural Networks, which are at the very heart of AI as electronic models of biological intelligence will continue to advance with the introduction of ANN chips.
The advent of Artificial General Intelligence, I believe, will happen in a few decades.