Whether it is at the beginning of your process or after some research and observation, write a list of assumptions. It can be assumptions you or users make, things you think would happen in some situations and actually don't.
I discovered this method in a course on Coursera: "User Experience: Research and Prototyping" that I would recommend. As an example they use a train ticket. To take a train, the assumption is that you need a train ticket, but is it true? What will happen if you don't have one, isn't there a fallback solution in case you were late or the machine was busy?
With this method you will easily realize which are the missunderstood points, the believes, what should be clearer and avoid missing something during the next phases.
At the beginning, you can make assumptions on the users: Who are they? Where does your product fit into their life? When, how is your product used? You can also make assumptions on business answering questions like: We will make money by? Our biggest product risk is? etc...
You can take 3-5mins per question in a workshop, then cluster the answers and create personas from it.
Check the Ebook from UXPin in the Ressources for more about assumptions (page 23-24).