I spent a couple years in the Content Management/Information Management circle using both FileNet and IBM BPM products. I would have to say I am incredibly unimpressed, and here is why:
1) BPM is one of those things I have not ever seen implemented intuitively. By that I mean (similar to what you say in your question) that a business user cannot sit down, and with 30 minutes create a process/workflow.
2) I have rarely seen a non-programmer be able to create acceptable solutions
3) The lack of ability to explain BPM to business users, the inability of business users to comprehend, and the amount of time and money spent on BPM seems ludicrous to me. From experience in everything I have been involved in, the company could hire 1 or 2 people (technical types) to work full time and create them a custom app for the time it takes a bunch of business types to have numerous meetings and hire consultants or train business people.
4) It is an incredibly niche market, and 90% of the outcome involves document routing and approvals. This is a VERY VERY simple concept, and one which a developer can usually code and implement a solution for much cheaper. If the developer is good, they can make it much more intuitive with less steps too.
5) It usually takes longer to have a coder talk with 10 people in management, a bunch of business users, learn to use a bad limited product and come up with a solution than it would be just to let normal development go on.
Don't get me wrong here - I know my experience, while quite deep compared to some, only deals with a couple products meaning I have only touched a little of what is out there. I am all in favor of innovation, but I haven't yet seen a halfway decent solution, and they all cost 20 times more than they are worth. I am sure COTS BPM type software will persist, but I hope I don't have to work with it for many more years. Hopefully then it will be halfway decent.