Pi isn't a panacea. Needs vary, so do solutions.
Pi is a better teaching platform for many purposes than
most of the alternatives, for loads of reasons that have
been done to death elsewhere. It's not even all about
the products - there is a momentum and a community behind
Pi of a kind which hasn't been seen for a while, and that
in itself adds something.
Some folks even find Pi products are an OK deployment
platform for some kinds of requirements, e.g. where
there is need for simple low cost readily available
physical IO (stuff which is readily available and
affordable in the Pi community and likely to stay so,
not buy ten milliona and we'll tell you about it, but
if you want another ten million in a year it'll likely
have changed by then).
Needs vary, so do solutions.
Having some understanding of what goes on beyond the
mouse can sometimes help people in understanding the
difference between appropriate and inappropriate - as
you've just demonstrated. What looks good on paper
may come with hidden snags, as I found when I did my
first ARM-based comms gateway project ages ago.
Based on StrongARM running MontaVista Linux, in an
industrial box maybe A5 size, it seemed to fit the bill
perfectly. Until we wanted to put our modified kernel
into flash where it could autoboot. The device vendor
didn't want to document how to do that, and there were
some challenges with the local IT department allowing
non-Wintel kit onto the corporate LAN.
Just the kind of thing that Pi might come up against,
except that sensible people nowadays will at least have
heard of Pi but may not have heard of MCUs or Beagles
or PCI/104 or whatever.
TLDR: Pi's good at a lot of things, and it's cheap
and fashionable. Just because something's fashionable
doesn't mean it fits everywhere, does it.