DB
Sorry but this has a lot to do with hand flying - and a lack of competency at it.
The original diversion from the desired climb out profile may well have been due to inappropriate use of autopilot coupling and particulary GA mode engagement, but that's not the reason this aircraft nearly hit the water.
That was because the pilot, mainly the captain flying, did not recognise the unusual attitude (excessive pitch up with decreasing AS) that prevailed for quite some time and take adequate measures to rectify it. Recovery from unusual attitudes must be something that a professional pilot can achieve. And that must ultimately be a hand flying skill. Ok sometimes pushing the GA button can help, but only sometimes, and subject to a good understanding of what other parameters must be established for that to work.
This smacks of the pilot almost continuing to bash the keyboard from the Children of Magenenta film whilst he should have, knowing that things were getting out of control, simply uncoupled completely, got the wings level, established an appropriate pitch attitude, power setting and airspeed, re-established a sensible climb. Then engage upper modes to stabilise thing, take stock, and continue.
Surely you can't argue about this? Or do you believe that big sophisticated helicopters have become too complex for pilots to be expected to be able to revert to basic PPL IMC techniques if the autopilot isn't doing what you wanted it to? Is there just too much panel info, making it hard to identify the basic instruments/displays to refer to?
It's easy to see how precious seconds can be lost trying to figure out why the aircraft isn't doing quite what you want when the AFCS is being used, when all the time the upset is worsening, and stress of trying to figure it out increasing. If hand flying skills were more current, I can't help but feel such pilots would be more ready to uncouple, recover by hand, get the aircraft flying again and then re-engage the AFCS carefully.