To avoid the "one platform" idea you will have to push the "best tool for the job in question" idea.
I'm not saying that "one platform" should not be used; but you need to use it where it fits.
If the company is really thinking about switching (which is different from some boss or manager ranting because he has had a bad day with his current tools or IT services), they should sit down and plan how they want to do it.
What is their core business? Will it benefit from switching to Linux (or another platform)? Does it make sense to also switch office and support tools to the same platform? (Developers are in a way easier; it usually makes sense to have them run on the same platform that your product or service runs on)
When they have a good idea about that, then it is time to sketch out a plan and start calculating the cost.
Remember, nothing is free. Even if you don't pay license fees for FreeBSD, Linux or some other FOSS, you still will have to pay for education / training and support.
If you can take away the fear and uncertainty, they only obstacle to selling your plan is the cost (ok, very simplified, but still).