Business intelligence for whatever a business does is largely unmet. I think it is hard to devote time to developing ideas (around mapping) that most business owners don't already know about.
You can pick just about any segment of a business and find a GIS answer for the person that does that job.
Probably just me blabbing but perhaps it is something useful.
Edit: As suggested in the comments I will try to make my answer clearer.
I often say that GIS (not for me but for the average person sitting at the desk down the hall) is a cloud. They may hear the term and may know it stands for gps (or they may know better), but don't know how it can apply to them. More and more people relate GIS to getting directions to Starbucks (which in itself is not bad). They do have a hard time understannding that practically every piece of information they have can or could be related to a position. Which means a GIS could be designed around that information.
With all that said on to the business of GIS in business.
My experience stems around... finding someone that will ask, talk to me, or otherwise show any interest in GIS/mapping. Then through visiting with that person find one thing about their job that has any kind of repetition to it and has an element that can be mapped. Whether it is customer zip codes (geocoding), drive time (routing), change detection (image processing), facilities management (enterprise gis), or any number of other business challenges. Find (especially to start with) something simple to show on a map and make it work for the (one) user. Expand from there!
One of the pressure points or holdbacks that business owners have is that everything relates to money in business. The cost of, and/or time lost are some of the first considerations they may have. Because GIS is still a cloud to them they don't see the direct benefit and first consider that their employees will be wasting time (this is past the point of spending money on the initial system) finding new routes to starbucks.
It amazes me at how little the cost to build a small system that answers a key business need is in this day and age. Software (osgeo, arcview, and others) is relatively inexpensive. Data can be acquired free or almost free and much of it online. Even hardware requirements are not as stringent as my early days.
With these factors in place there should be and really to some degree is, a boon in GIS related services. A good programmer with some GIS savvy and good people skills would be able to coax smart, adventurous CEOs to listen.
I think that also it might be possible to pull an inside job and utilize open source data and software to build a "free" system that fixes a problem and get the attention of the CEO (Do it first and ask forgiveness or ask for funding for open source project).
To keep it all from sounding so simple: There is a requirement for all GIS systems. That is validity. What I mean is that the data must be touched, cleaned, updated, and maintained from now on (constantly). It is a good idea to count the cost and design the maintanence system at the beginning (not as an afterthought). If users can't trust the data they won't use it.
Use the KISS (keep it simple stupid) method. It is easier today than it has ever been. Only 5 years ago it was relatively unheard of to expect a national force of GIS people editing your data. Now with OSM (open street map) that is exactly what is happening (especially if your business requirement is national)! So using KISS you can have a few datasets that you spend most of your time maintaining. Using online or regularly download offline data for background and services. Serve it out to the web and desktops with opensource. Using COTS software for the heavy load data analysis and manipulation.
What a dream job!