This is a tough problem for companies. At my last job there was an "advanced concepts" group which had a lot of freedom and latitude, but they had a poor reputation for not really producing anything useful. The few things they did produce had to basically be redone since the quality of engineering wasn't ready for production.

I'm not sure how to fix that, since lack of freedom wasn't the problem! It probably is just making sure that the right people are leading.

I also don't know how to reduce other engineering groups more focused on practical projects resenting the innovation group working on more fun, low stress projects. That is a hard political problem.