Thinking mobile first for enterprise software might help address this situation. It is possible to think about features and functionality for mobile personal devices such as iPads and smart phones first, test them out with a few users, prove the need and build such functionality on a larger scale and on the web later.
For a example a custom application for a particular role or audience can be built with an investment of a $ 100,000, assuming the basic objects necessary to build the mobile app are already available. The feedback from such a mobile app can then be used to justify and thus prioritize the development of the web version of the product.
In many cases nimble partners might be able to share the risk with product managers to experiment, if they see long term benefit. Such an approach makes innovation and experimentation possible even in large enterprise software companies or installations.