As engines has stated, the SAS is far more than an autopilot. It has varialbe gains for different parts of the envelope and compensaion terms for flaps power airbrake etc.
Variable gains are nothing special in an autopilot; it would be next-to-impossible to have a single set of "inner loop" gains on a modern civil airliner autopilot and be able to use it to achieve both RVSM accuracy in high speed cruise and CAT II/III landing approval. The "aircraft gains" resulting from the very different aerodynamic control effectiveness at such different speeds can only be accurately controlled with system gains that are suitably aligned.
To be frank, what's described (unless it has full control authority, which seems unlikely) is little different to many civil certified systems in use today, and probably not too different to some systems which have been around 20-30 years or more - I believe the L-1011 had/has a system with at least some if not all of those features....
Are the actual engineering development / flight test teams split, or are these all at Warton?
I believe they were split at one point (early on); whether they have been reunited since, I don't know. (But it's not a great idea to have the design teams removed from the final line either - there's a lot of engineering support required during build, especially when the build run is so short that you'll have finished the run before you work the kinks out of the build...)