Laughable: "It's a business site, I can't use CSS"
That was the quote of the day in #css on EFNET. If anything, that is just the opposite of the case. Business, far and away have more reasons to consider CSS than some crappy personal site (not all personal sites are crappy, but most are).
The first concern is almost entirely aesthetic. The comment usually goes, "my client wants it to look the same in all browsers." I'm sorry, but what a load of crap. What your client is trying to say is that, "in browsers that most people use, it should look good." Of course it should look good. As a web designer, if you handed a text-only, unstyled demonstration to your client, they'd walk out the door.
What is less obvious about the client's request is that "all browsers" sure as heck is more encompassing than they realize. Realizing that table designs suck on pdas, cellphones, internet appliance, text browsers, and anything older than 4-5.0 browsers is something they're not accustomed to thinking about. That's your job.
I'll assume for a minute that not everyone is so articulate that they could make the slam-dunk argument from aesthetics alone. There are pocketbook reasons why businesses should care.
First, up front design costs are trivial compared to the ongoing cost of maintaining a good site. Most savvy organizations and businesses know that from experience. Their experience might not come from the web, in particular, but from other software vendors. It's never the initial purchase that hurts, it's the ongoing maintenance costs (in software it's usually about 20% of the initial cost).
Maintaining, upgrading or updating a web site is no trivial task with a table design. The complexity of the code often prevents a client from maintaining their own work, which is really nothing but job security for web developers. Furthermore, upon redesign or upgrade, it often means starting from scratch. This is not 10-50% maintenance cost, it's 100% of the original.
Second, unlike most personal sites, businesses are interested in conveying as much information to as many people, as quickly as possible, and for as little as possible. Let's imagine that about 1/10 people have an impairment that makes using the web very difficult. A table layout is almost inherently inaccessible, or even where the developer has taken care to meet the requirements, the layout may be difficult to navigate without the visual cues. As such, that's potentially 10% of the market the you've foreclosed on simply because.
The accessibility issue plays on the lawsuitaphobia of many organizations and corporations. Fear that someone, somewhere might file a discrimination suit will probably scare the bajeezus out of many bosses. This argument is probably better made in larger corporations with higher exposure.
Third, while the initial design might incorporate the original content, inevitably, the content will grow with time. With each new file, presentation is replicated and pushed back to the server. Over time, this increases the bandwidth requirements and server requirements mulitudes faster than separating the content and presentation initially. Here again, the maintenance issue rears its head too. When the new documents are created, they're often embedded in the templates, extracting that content later is time consuming and costly.
Finally, since corporations and organiations, unlike personal sites, are usually looking to invest into the future and are in the game for the longer haul, higher initial up front costs are quickly out-weighed by the countervailing requirements of the future.