IIMC is a constant threat in the mountains in poor weather, day or night, which is why the use of escape headings is important and having a plan before you commit to the inside of a feature is equally important.
They clearly did have a plan which was a 180 out of the bowl and down the valley - something they had executed well in their previous attempts to get to the climbers.
The weather does change rapidly in such terrain and getting caught out is always possible - there is a often a second or two before you acknowledge it isn't just a little bit of random mist or low cloud that you won't pop out the other side of and that you need to take action.
Perhaps this is what happened halfway through the valley/bowl turn but it appears they climbed which may be cause or effect, only they know for sure.
Whatever the reason, they ended up in a very scary position, IMC inside a mountain feature and having to rapidly transfer from NVIS to instruments (not easy in the best of conditions).
I understand the reasons for trying to use the AFCS upper modes to help but they became a further distraction and prevented a roll out on the escape heading. Thereafter the only solution was wings level and pull power - which essentially is what they did.
ODEN - in answer to your first question they were trying to get to the climbers and would have identified the bowl feature as accessible before they went in - I presume the intention was to achieve a hover inside the feature and then assess the feasibility of a further hover-taxi up the side of the mountain.
As I mentioned before, the Bristow training team clearly thought this was a scenario worth repeating in the sim for other crews to learn from. Pilots make mistakes - it's human nature - doesn't make them bad pilots and the best pilots I know learned from their and others mistakes - you just have to be bold and honest in identifying the error and the prevention.