Technology has a viable lifetime before you need new technology. As a technology company, you therefore need to be invested in creating new technology or be willing to go in to the spiral of customer loss and only dealing with legacy systems before a slow death.
Statistically speaking, if you know most of your staff is intelligent and has experience in your market, you're more likely to get an outlier idea (one that is way off on the end of the bell curve, ie, actually really good) by casting a wide net, and listening to the bulk of your employees.
The problem is that to go anywhere, ideas need time to gestate and develop, so you have to give all those employees a little bit of time to develop new ideas for your company.
It's been a while since undergrad, but if you want, I could try to bust out an equation modeling (and predicting the expected value of) the likelihood you hit a really important idea in the wide-net situation versus the "dedicated research staff" one.