It is hard to argue that AF would not bring value to an organization, but the barrier to entry may be very high. For a company that already has a well-established point based PI system, lots of value is likely already being derived. How can you justify a large infrastructure project to turn the point based system into an AF/hierarchical model based system?
Infrastructure projects are always difficult to sell. They provide the mechanism to do lots of good and valuable things, but by themselves, they provide new capabilities and no immediate perceivable value. As a result, it's important to tie an infrastructure project to an application or specific business need. Of course, no one application is likely to require the full AF structure, but it will require some portion of the hierarchy and/or attributes. So, the key is to build a reasonable complete hierarchy without concerning yourself with including every attribute. You see, one of the nice things about building an AF structure (especially if you use templates) is that you don't have to get it right the first time. As you find additional attributes needed, they can be easily added. If the attributes are added to templates, then every instance of that template receives the new attribute automatically. In addition to adding attributes, you can add calculations (AF Analytics) giving you more flexibility and capability.
One of the really great things about AF is that you don't have to get it right or even complete on the first go-around. It's easy to add attributes, analytics, and notification at a later time, especially at the template level. When adding components to an existing template, all of the elements based on those templates will automatically become updated once the changes to the templates are checked in.