The first formal step in any project design process should always be to think about audience, or users. This is one of those questions that seems fairly obvious, but can end up being more complex than expected. When working with faculty, a PI may state that a project is "for" an intended audience, but in the course of conversation, it becomes apparent that they have someone entirely different in mind.
Whenever possible, encourage collaborators to think about audiences plurally. In other words, it's good to consider multiple types of people who may be interested in a project (community members AND academics, or activists AND students). There is only one answer that is unacceptable for the question of audience: everyone.
It is also important, when considering audience, to consider whom we are talking to and whom we are talking about. In other words, if a project or study contains information about a particular population of people, it is ethically important to ensure that this population is included in our list of intended audiences, and that we are accounting for how they can access the project. If a PI has not considered this, we need to make sure that they begin thinking about it and understand why it is important.
In general, I would recommend identifying one primary audience and up to two additional audiences.
Identifying project purpose often takes place at the same time as identifying audience (See the above question: "What is the goal of this project?"). Bear in mind that projects may have multiple purposes for different audiences; a website that serves as a companion to a history monograph may be used for one purpose in an undergraduate classroom, another by a peer scholar, and yet another by an activist. It is once again helpful to identify a primary purpose, as well as a few secondary purposes.
For academic projects, it can sometimes be hard to orient ourselves to the idea of a "project purpose" at first. We may be used to thinking of things as "the goal is to prove x" or "the purpose is to show that y." Reframing this around a particular audience or context can help: to whom are you proving x? Who needs to see y?
There are some design methods that can help re-orient collaborators to consider the purpose of their project, and for whom it might serve that purpose. Below are exercises that can be done with PIs to help them picture the "who, what, why" of their project.
You can choose one of these, or mix and match elements of multiple to make a customized design workshop for your collaborators. All these methods (and more, and customized versions of them!) can be useful in different circumstances - the trick is finding the right method for your use case. If none of these quite fit your needs, I recommend Universal Methods of Design: 100 Ways to Research Complex Problems, Develop Innovative Ideas, and Design Effective Solutions. It's a good set of resources and inspiration for design research methods.