BI is Not an IT Activity
The gulf between the information needs of real human people and IT is almost universally experienced. It's helpful to recognize the drivers for this situation in order to attempt to correct it. Some of these have been mentioned in previous posts.
There's a fundamental, unrecognized, nearly fatal flaw in the ointment, revealed in the term "IT" itself: the casting of information acquisition, management, understanding, use, and communication as a technological activity, hence "Information Technology".
Information is a human property. Casting it into the realm of technology presupposes that it's a physical resource that can be effectively processed with an industrial assembly line approach, given the acquisition of the appropriate machinery, the installation of an adequate industrial-production control model of oversight and management, and staffing up with the appropriate level of resources (not people, revealingly). Once everything's set up, the big green "Go" button can be pressed (metaphorically) and he wheels will turn, the gears will mesh, the data will be pumped in, and the information the business decision makers need will come out.
If you've been underserved by your IT department in getting your questions answered, your information needs satisfied, almost everything you've experienced, all the trouble getting results from IT, is caused by the application of this paradigm. And almost all the solutions suggested are applications of more of the same to address the weakness inherent in the model.
One example: the creation of a Business Intelligence Center of Competency (BICC) as the solution to the problem. It can work, but only if it's not just the addition of another mechanical/operational command-and-control structure with the intention of assuaging the pain causing by the existing environment.
In reality, too many BICC initiatives start out well, then crumble into the same industrial dust as the IT departments whose functions they're replacing. As soon as friction sets in, and it will, usually in the form of someone important not getting something they want as soon as they want it, the impulse is to add mode control systems to manage the submission of requests, the analysis required to understand them, and the prioritization and queuing systems to manage their satisfaction.
All of this misses the essential, central truth: business intelligence - the analysis of business data in order to achieve information and insights from it, and to understand the stories it has to tell, are human cognitive and intellectual skills.
It's important enough to repeat: BI requires human thinking skills. It's not an industrial process.
Business people who need to understand their own data understand this, which is why they use Excel, Access, and increasingly, the new (now 5+ years old) generation of direct-access, low-friction highly effective data analysis and visualization tools to help them understand their data. And they're increasingly frustrated with their IT organizations, which don't get it.
Recasting the concept of BI into its proper form is an essential element in solving this problem. Layering on more industrial management processes only compounds it.
Successful BI occurs when it's recognized as a professional practice, requiring of its practitioners the appropriate combination of aptitude, interest, knowledge, training, and experience. Then the barriers can come down and the myriad distances between real human people and the business intelligence locked up in their data can be freed.