Gabor's posts here have indicated that they're doing stuff with NLP and machine-learning. There's a lot more you can do with that than what we'd think of as a simple Outlook plugin. Think of a smart personal assistant that understands each email, can file and classify them, trash the spam, send out canned responses to common queries, and flag your cell-phone or PDA for urgent messages from trusted folks. It could replace secretaries.
That, and investors are likely the people who have the most need for such a tool. So it'd naturally look better to investors than it would to the rest of us. That's the risk of putting all your ideas before a certain group of people.
> Is the better growth strategy to put something in front of investors first and see if you can hold off on the public launch?
I doubt it. Historically, that approach has had very few successes and many, many failures. Once in a while you may get an Amazon.com or an Akamai that takes tons of capital to get started but then has a wide moat and enormous profits. But more likely you'll end up with a Go or a Value America or any one of a number of dot-com flameouts. (Even that's misleading, as Amazon got users and went profitable before taking outside investment, then operated in the red for like 7 years while they built up infrastructure.)
Personally, I'd put Xobni at the top of the YCombinator deadpool, with Loopt in the top quarter, even though those seem to be some of PG & investor's favorites. But then, they're swinging for the fences. If they succeed, they'll also be among the most lucrative.
(Incidentally, I'd put Wufoo, VirtualMin, and DropBox at the bottom of the deadpool, i.e. most likely to succeed. You may all publicly pillory my predictions later...)