The idea, is solid, so where it starts to count is the execution. On that topics, my advice is to start releasing working code as a hosted service right away. Large companies can talk about mashup servers, and other Web 2.0 ideas, all day long, but until they not only show the "public" those servers in action, and study the way people start using and wanting to use the system, they won't find the quick traction a movie target like Web 2.0 technologies need to be more than a passing fad behind-the-firewall.
Indeed, releasing working code as soon as possible -- the permanent beta -- is one of the keystones of all the Web 2.0 technologies out there. The success of the Web 2.0 All-stars -- Google, flickr, and del.icio.us -- is created in large part by simply getting their code out there and used by 1,000's of people.
Don't Let the Firewall Hit You On the Way Out
As Koranteng and I discussed over post-BBQ tea and coffee when we last met, the notion that there's a difference between the way people want to, can, and will use intranet software and internet software is crap. The idea that the software used behind-the-firewall has to be different than that used beyond-the-firewall is a pleasant fiction the industry makes up.
That line of thinking is strongest when it comes to crop-dusting enterprise software with Web 2.0 technologies and thinking. And a major part of that realization is that IBM and all the other "enterprise mashup server" projects need to get their stuff up and running on the public internet double post-haste. Crowd-sourcing is the way to make Goose 2.0 lay endless golden eggs, and you'll never get a valuable enough crowd by taking the traditional behind-the-firewall, silo-bound approach.