So, you can use something like this to determine what they might automatically know/be able to do without rolling. Though, for certain difficulty levels (Hard+) I don't think you should ever not roll, by definition they are risky and should carry that risk when performed.
This is completely context-dependent. If you try to pick a lock, for instance, the risk is not in the lock difficulty itself, but in the chance of being discovered (and more accurately, of being discovered while picking it). If there is no chance of being discovered, then it's a fairly boring yes/no proposition and you can avoid a roll - if the character *could* do it, they will be able to.
On the other hand, if there *is* a chance of active consequences, rolling is interesting, but the interpretation of the result is even more interesting. A failure might not be indicative of the lock not being picked. Rather, it might be indicative of someone noticing the picked lock, either while it is being picked, or after it has been picked and the characters have gone through the door (and hopefully closed it behind them). Or an incoming patrol forces them to find a spot to hide, leaving the lock sort of half-picked (and maybe the picks still inside the lock...).
The same logic can safely be applied to most actions that are not actively resisted : if the only consequence is that the intended benefit is not obtained and there is no personal risk, then the yes/no proposition can be resolved without rolling: the character tries their best, and their best is either enough or not. Failure is interesting when it makes the lives of the characters more interesting. An Investigation roll is interesting in what happens in case of failure: does the PC attract attention ? Do they ask too much carelessly and trigger people who don't like busybodies ? Do they draw the attention of the watch ? Of rivals ? They might get the information they seek, but be unable to do anything useful with it because they inadvertently triggered events that cancel its usefulness.
Rolling is not an end in itself. If you don't need the "more interesting" part, then you can safely skip it and use a 'best effort' result.
Of course, if there is active opposition, then all bets are off.