I'm trying to figure out who the intended buyers for that pc are, and I can't
Amstrad had largely been shoved out of the low-end home PC market by competition from Commodore and Sinclair,
This isn't true at all. Amstrad actually entered and already-crowded marketplace and quickly rose to be the third-biggest platform in the UK. They actually squeezed Sinclair out of the market. Commodore 64 and Spectrum launched in 1982 and had a huge head start. Amstrad launched in 1984, quickly rose to the #3 seller*, and by 1986 Sinclair ended up selling the entire Spectrum brand to Amstrad. So Amstrad came from nowhere and ended the 8-bit wars with about 70% market share in UK, and a lot of Europe, since they actually ended up owning the Spectrum brand as well. The Spectrum was then redesigned using Amstrad's design philosophy. See the Spectrum Plus 2.
Where they hit problems was in transitioning into 16/32 bit. But that's a common thing. Most companies failed this hurdle.
EDIT: * note the reason Amstrad succeeded in a market crowded with dozens of offerings. They were the masters of "bundling" and actually pioneered all-in-one stereos. They'd make "expensive looking" combined stereos, so you'd have a music system that looked like you had bought stacks of audio gear, but it was way easy to set up compared to the real thing, and much cheaper.
The Amstrad computer was designed similarly. The head of Amstrad noted that current low-cost computers were unprofessional looking things that you had to plug into a ton of peripherals and a TV. So you got shaky video out, fiddly wires, and you couldn't be using the device while the TV is in use. Amstrad realized they could bundle some components, stick a dedicated monitor on it, and design it with a more professional looking keyboard (Spectrum keyboards looked like a big calculator). So, you could game on it, it looked better due to crisp RGB out, and you could use it as a dedicated word processor in an office, much cheaper than shelling out for an IBM. There's no way you could hook up a commodore or spectrum to a TV in an office and expect people to use business apps on those.
EDIT2: Note, this is where later efforts such as the "Amstrad Mega PC" failed badly. While it had the "bundling" thing down right, where it failed horribly was in that it bundled things that had no reason to be together, and it completely failed where earlier Amstrad devices has succeeded: in understanding the normal person's use-case for different gear. They took the
complexity out of setting up stereo component systems, or the complexity out of hooking up a home computer, while also doubling as a super-cheap alternative to either hifi component systems or a proper word processor. Sticking a Sega Genesis into an IBM PC doesn't meet those criteria. It's like bundling a toaster and and alarm clock, it might be cheaper than buying them separately, but there's no use case for that (why would you want to play Sega games at a desk?).