Thanks to @OFSROW for their tweet with a nice graph visualizing the different traits of AI (click to see the graph). And here is another excellent compilation of AI concepts and terminology (clik to read).
It can be confusing for a compnay wondering whether or not AI is for them to understand the difference between machine learning, NLP (natural language processing), vision, robotics, and expert systems.
Even more importantly, it is confusing to figure out what solutions are "real" artificial intelligence, and which are algorithms pretending to be artificial intelligence. Not that algorithms would not be useful - but if they are written by humans, they are not artificial intelligence and instead of learningm they have to be rewritten when things change.
So to oversimplify a bit: if you need a solution which can learn about data and adapt accordingly, choose a learning system. If you need to find causalities, which will remain the same for a longer time, use algorithms and remember to do analytics regularly to see if things have changed so you can rewrite you algorithms.