I was just quickly thinking of examples. Pick lock, possibly not help-able (although two players of rogues with expertise in thieves tools might well disagree!) But many other tasks, likely yes. Of course this will always be up to a DM - one could just squash helping.First, why are you allowing advantage on a lock pick. That's not a helpable action in my games.
Ah, so to get a "nearly impossible" result in tier-2 with +10 and advantage, is about 9% likely*. To get a "nearly impossible" result in tier-2 with +10, guidance and advantage, is about 41% likely. The cantrip is making quite a difference. Even without help, the change in getting a "nearly impossible" result is from 5% to about 18%.3 prof + 3 expertise + 4 dex + d8 bardic inspiration = 12-38. I don't see that as a meaningfully different range than 13-42
IMO this shows that it's actually not guidance that's the problem!
So doing the same exercise without bardic inspiration because we both agree it's a resource and should be able to have strong effects.
Without guidance = 11-30
With guidance = 12-34
However, it's also important to respect ability spiking via stacked buffs (which led 3rd ed to have that whole buff-types thing, and systems like Earthdawn to apply a rule of three). Guidance, being nigh-costless, is egregious in that it freely pushes everything upward. Above I think where the designers - from the extant text - appeared to have expected the limits to be.