Point at a bright star and slap the Bahtinov mask on.
Set APT to short exposure, max gain, 1:1 (sometimes even add the magnifier if the nearest bright star isn't all that bright) and very gently manually focus the lens.
When it looks good, I take a longer sub so I can really see the diffraction spikes well and verify they are centered.
REMOVE BAHTINOV MASK
Then back to the target and resume the imaging plan.
Takes less than 5 minutes. (I also plate solve/sync/goto/platesolve/display in Stellarium again when I get back to the target so that is part of the 5 minutes.The CEM 60 is rarely very far off, but I'm OCD.)
APT has an autofocus routine (obviously worthless for Nikon glass) but I haven't been able to get it to work better then the Bahtinov yet on my Edge 8.
I am just setting up an AT115EDT (added the Pegasus FocusCube over the weekend) so maybe that will work better with APT??