Mike -
just for curiosity's sake, I pulled my early generation DS1000Z from the basement and hooked it up to my "improved" DG811 (~DG992) to test direct signal transfer. It appears to be working quite well up to a sample size of ~1MSa from Scope memory. But the system is quite finnicky vs. the settings so you may need to experiment a little until you get it right. It's important that the scope doesn't display a higher voltage than the AWG is capable of outputting, so if required, you may need to enter a probe scaling factor to keep the dsiplayed voltage at less than 20Vpp (hi-Z on the AWG). If you get a warning message during import to the AWG ("signal level exceeded" or the like), chances are that the whole signal will be corrupted.
When entering this "signal copy" mode, the AWG will automatically change to "Sequence" mode, and this always affects both of its channels. You still can use the other channel to output standard waveforms, but you need to configure them as part of its "sequence enseble".
Moreover, at larger sample sizes the waveforms appear to turn a little awkward, don't know what's exactly going on here. For testing, I used a 1Mbps PRBS signal that obiously contained many slopes in the recorded interval. I got "some" waveform even at 6MSa, but it clearly differed a lot from the original. Maybe the AWG is performing some decimation at these large memory chunks, though as per its specs, it should be capable of up to 16Mpts. Whatsoever, I think this combination should work for you.
Cheers,
Thomas