Okay, you're at a regional office party for your Fortune 5000 employer and a really cute user manager has cornered you after one drink too many and asked for your advice on how to beat IT at its own game. What do you say?
1 - find and adopt a guru. Everybody in your position has a secret (at least from Corporate) Wintel support guy or organization on call -and sure they come quickly, mark their invoices "Garbage Removal" or whatever you want, and generally get along great with your staff; but how do you know they actually have a clue?
The answer is that you don't - and their willingness to support Windows speaks well of their business sense but doesn't say much for their expertise.
So the most important thing you need to do is to find and adopt a guru - someone whose technical knowledge is real and whose opinions you can trust.
How? talk to people you know, send a lieutenant to school, recruit from the science faculty at the local university - it's a trial and error process much like dating. You need to find someone you can trust to interpret what you need in terms of technology that works, to help you argue your side in the corporate IT wars, and above all to tell you when you're wrong before you get into trouble.
2 - Most of what the IT community cares about is BS from your perspective - but there are trends and technologies whose implications, if understood and applied, could directly improve your bottom line. Create a little group of kitchen IT cronies, bring your guru into the picture as fully as possible, and meet once a month or so to bluesky ideas.
The latest whizzybang from Intel or AMD is just that: the latest whizzybang; but Sun's CMT and IBM's cell are truly revolutionary - similarly the latest hot open source super duper file system may be exciting stuff to a techie - but Sun's ZFS signals a whole new ball game in information storage and retrieval. Blackberries and Treos are more of what's sold before - but the iPhone is the beginning of a whole new idea. Are there implications in these things for you? IT isn't going to tell you - so form the right group, kick things around, and take the time to find out for yourself.
3 - the more technology bigoted and vendor centric corporate IT is, the easier they are to manipulate. They want you to use AIX and Global Services for everything? Greet this idiocy with enthusiasm - because the combination of multi-year planning horizons and itemised, daily, billings on last year's panic projects provide perfect cover for doing your own thing in the niches corporate IT doesn't care about and never gets to.
4 - It's true that forgiveness is easier to get than prospective budget authorisations - but the best thing of all is interim funding from off (IT) budget sources. Pay for thing out of other funds, justify it to corporate as interim -crudely meeting the need while the professionals at corporate IT do the evaluating, planning, budgeting, and consultant hiring needed to get it right. Sure that's a waste, but it's not peace at any price, it's peace at their cost - and meanwhile you get whatever benefit the thing provides to your business plus the political benefit of team playing an appropriately subservient role with central IT.
Remember that what counts for central IT is their mandate, their budget, their director's span of control - and he's probably no more interested in technology than you are; but he has to pretend and you don't. So use that: if his processes revolve on long lead times and are held in place by vendor reliance, leverage those things against him: do things now and sell that as buying him time to get it right.
5 - Never get into development or sweat technology differences: build an appliance computing mentalitity among your own staff. Brand X vs brand Y matters to technology experts, but not to you. For anything you put a minute or a nickel into, the only thing that matters is that it work - quietly, in the background - like a telephone or a good employee. Bottom line: if it gets in your face every day, or even every couple of weeks, you're better off making it somebody else's problem - hello? corporate IT?
6 - it's sad but true: generally speaking technology quality and popularity are inversely correlated for IT. Macs are genuinely better than PCs on every dimension except popularity, Linux is genuinely better than Windows on every dimension except popularity, Sun's UltraSPARC products are genuinely better than their x86 products on every dimension except popularity, etc etc -but remember that the reverse isn't true: unpopularity does not always indicate quality, a few popular products, particularly services products, are actually quite decent.
Sometimes the appropriateness, or otherwise, is just not obvious - as I said, it takes a guru - but remember that the rule applies there too: someone whose opinions are generally with the majority is generally wrong.
Oh, and remember that I never said a thing - wasn't even here.