The pointing of the telescope was not perfect. We set the pointing by calibrating the mount on a target field of stars the position of which we found using image-analysis software (see astrometry.net). The centre of field was then added, by hand, to a setup menu, and the telescope would centre well for a while. After a meridian flip, or after a few days, the centring was not so good any more.
We tested polar alignment by the drift method, but results were not clear. We are thus unsure why the pointing of the mount would deteriorate.
GOTO telescopes have calibration modes where a sequence of observations at different parts of the sky are made so that a pointing model can be calculated and used. We could not do this - we did one pointing and updated the model based on that - we needed a way to access the 'build a pointing model' system by software. Even to do this automatically.
We had the basic ingredients of a system to take pictures of the Moon and centre accordingly, but it was not functioning robustly.
The dream solution would be something like this: the system is lost and does not know where it points - so it takes a picture of the sky and gets an astrometric solution - then it gets more - if one solution does not work it gets more pictures, at random
above-horizon, places - and keeps going until a solution is found. Then it goes about its work. Sigh.