He notes that what should be interesting is the software infrastructure required there. And in particular, what connection it might have with SocialSoftware.
His take, that enterprise software is not good at handling exceptions - out of the ordinary events and requirements, and yet this is what's needed. Software enabled systems need to be able to ping humans to inform them of exceptions, they need to help humans track down and talk with other humans who have the right knowledge and responsibility to handle the exception, to freeze the context within which the exception occured so that people can analyse and explain it, and to enable the rapid coordination of people to help find a solution.
JoelSpolsky (who makes BugTrackers) finds a software ExceptionHandling problem getting escalated to an organizational one : http://www.joelonsoftware.com/items/2006/05/05.html
One of the hallmarks of a broken system is when there's just no possible way that the programmer who is writing code that talks to customers can ever get feedback from those customers about bugs, because the call center is outsourced to a different company than the software development project is outsourced to. Everyone is trying their hardest to do their job but management has set it up so that it's impossible. (OutSourcingAsAModularityMistake)
This is gonna be one of those things that happens as business (organizational design and management) and IT become even more intertwined. We need ways for the software designers to get feedback from the organizational architecture.
JoelSpolsky on customer service : http://www.joelonsoftware.com/articles/customerservice.html
Thingamy's Sig getting to grips with something similar : http://thingamy.typepad.com/sigsblog/2007/03/businessproces.html
See also :