Just realised that Beth Richard should be able to shed some more light on the CD1200. Her name is on the silkscreen and she was heavily involved in the CD32 and FMV projects at Commodore.
Did you not read the whole of the thread?
http://www.amiga.org/forums/showpost.php?p=816409&postcount=25Also I think that there might have been problems with Akiko and the 16-bit PCMCIA interface. Not that Akiko was particularly useful anyway.
It's not essential as the games went through the OS, which used the CPU if there was no akiko. All of the PCMCIA CD solutions that came out relied on that for compatibility. The CPU routines weren't the most optimal, so you really needed an accelerator. There were patches later on that improved them, but I don't think the 020EC in the A1200 could ever truly beat Akiko. Some hardware on the PCMCIA slot should have been able to do it quicker (it's a pity they didn't go ahead with the plan of chunky pixels for AGA).
I think at the point the CD1200 was done then commodore were really scraping hardware together from whatever they could find and they already had akiko from CD32 so they just used some glue logic.