Rework Windows 7 to be touch-proficient. This won't happen by the expected fall release of Windows 8, but if Microsoft really believes that Windows 7 should live beyond legacy, it needs to make that OS touch-savvy -- and not that crappy "Windows for Pen" environment it's been trying to foist on users since Windows XP. Touch means using fingers, not pens. The truth is that, on touch devices, Windows 7 is a bad UI -- the icons, menus, and other UI elements are too small to touch.
The place to start is to have contextual DLLs for the major UI components that most apps use; the touch-savvy UI appears on a touch device by default, and users can switch the display mode if they're using a mouse or trackpad. Windows 7 and its applications should be at least usable via touch gestures if they're to run on touchscreen devices.
I realize this approach is clumsy and will mean work by app developers, but it's more elegant than the poor experience Windows 7 now offers. Of course, the real answer is to abandon Windows 7 as soon as possible. Realistically, that will take years, and if Microsoft insists on carrying Windows 7 into the modern world, it needs to modernize Windows 7 enough to work plausibly. Apple has done a good job of this in Mac OS X Lion and OS X Lion, so use that transition as a guide.
Get rid of Metro's x86/ARM distinction. It's nuts that WOA isn't equivalent to WOX. The reason no doubt is that WOX relies in legacy Windows for native Windows client management, signaling a failure at Microsoft in its effort to port that technology to the ARM processor. WOA shouldn't ship as a second-class Windows Metro. In fact, users should neither know nor care whether their Metro tablet uses an x86 CPU (from Intel or AMD) or an ARM CPU. Neither should IT. That means WOX and WOA need to have real apps and be equal in all respects.
If users must lug around Windows 7 on their tablets "just in case," they can buy x86 tablets that have both Metro and Windows 7 Plus. Likewise, if they buy into the notion of convertibles -- laptops whose screens detach to become stand-alone tablets -- they'll get a WOX device. But a pure Metro tablet should have the same capabilities regardless of the CPU architecture.
Make Metro a strong OS in its own right. Microsoft seems to be treating Metro as a UI overlay to Windows 7 on all but ARM devices. That makes it too easy to rely on Windows 7 as a crutch that lets Metro stay underpowered. Imagine if Apple had put Mac OS X on the iPad, as many pundits demanded: App developers wouldn't have bothered creating compelling native iOS apps. But Apple forced them to do so by providing no such crutch. The fact that all Windows devices but ARM-based ones can run Windows 7 (albeit poorly) means that Windows developers will be slow to take advantage of Metro -- and will simply perpetuate the Windows legacy.
Even if Microsoft separates the two OSes as I suggest, that Windows 7 legacy crutch remains and will slow the development of strong native Metro apps. To combat that laziness, Microsoft needs to make Office 15 for Metro a truly compelling app. It needs to be at least as capable as Quickoffice or iWork on an iPad. To do less is to proclaim that Metro tablets are fancy e-readers -- niche products like the Amazon.com Kindle Fire, not true computing platforms like the iPad.