Hehe, it depends on the position, the analysis method that may find the best move in some position very quick may not find it ever on another. This is mainly true for the differences of the goals on the position:

What are you trying to do? Are you defending a dangerous game?, Are you on the attack on critical positions that could lead to a won game? Are you in a quiet position where there's nothing to do and you need ideas of how to continue? Is the position double edged or sharp and finding the best moves is difficult?

All these situations and more, require different methods of analysis, there's no win button and you may find yourself changing your goals as the game progresses.

For instance, to win with the black pieces you start with a disadvantage trying to equalize, then after equality, if it's still early in the game, you may want to try a complicating strategy in the hopes that the opponent will slip, or just play solidly if you think he'll overreach, then after gaining an edge on the game, playing to increase it by going into dangerous positions for the opponent, and finally, securing the win, making sure that you don't give it away by some defensive resource of the opponent, or trying to reach a position where the winning moves are easy to find.

The method of "going into dangerous positions for the opponent" may suck if you haven't reached equality yet, so before trying to find the best move you have to understand in what quadrant you're in. Usually this is told by the engines, they show a side with advantage when they have a score of about >0.40, near equality by <0.20, and winning positions (not "won") for around >0.80.

But this is not the full story, there are several positions where the engines are clueless, positions in where they think black has an edge but white is actually winning, or positions with marked 0.00s (the engine's way of saying they don't see more than a draw) but in where one of the sides could easily lose by playing natural looking moves. A good strategy is actually go for those positions, if you have a better understanding on them than engines.

I suggest you get used to stock analysis methods before trying more advanced ones:

1- "Jump" Analysis: The most basic interaction, you just let the engine analyze a position for a given time, and then play into the board its chosen move, and take note of score changes. This can be done until a set of moves or until a threshold is reached, for instance, if the initial score is 0.20, you may decide to force moves until it goes up to 0.30 or 0.10. What is important is that you see what moves are played and see if you like the engine's plan.

If the engine goes to 0.30 but black has strong attack against your king after sacrificing a piece, you may want to avoid this variation for being unnecessarily complex. You can also see to where the position is going with this engine, if white gets higher scores easily it can mean that white is better than it looks, or that the engine has no clue of how black should defend.

2- Wide Analysis: This one doesn't force moves into the position, but rather asks the engine what moves it would play if the main move wasn't available, and then what move it would play if THAT second choice wasn't available, and so on, by using the searchmoves or MultiPV features of the engine. Very often the engine is not going to pick the best move at a given short while (if they did the OP question would be moot), but they do when you exclude the main moves until the best one "pops up", for easy positions, at least.

Once you are happy with the alternative moves of the position you can start forcing them into the board, but as you go along, you also exclude the best replies to the moves you force, so at least near the root you are unlikely to miss strong positional moves. And don't forget to force interesting moves that don't look as bad, if after excluding the main move the engine says the second best is to hang your Rook, but with a score very similar, it's likely a critical variation.

While method 1 suffers from tunnel vision, the second one is very weak tactically, but it may be effective to mix them with discretion.

3- Crosschecking ideas with different engines. This is very easy, you just use several engines on the position, and check the ideas that one disagrees with, with the other. If engine A likes e4, and engine B likes d4, you force d4 into engine A, and e4 into engine B, and then give each engine their color (in this case, for e4, engine A with White and engine B with Black, and for d4, engine B with white and engine A with black), and force their moves into each other, as in "Jump analysis".

Eventually either Engine A will convince B that e4 is better, or engine B will convince A that d4 is better. Besides looking at a playable line (strong enough to refute one of the engines) the user also knows which engine has a better understanding of the position for this side (that an engine is the best finding the best white moves on the position doesn't mean that it's the best on the black side too.)

4- Human input. While the above methods can be simulated by an algorithm, this one cannot. After using methods similar to the above, which line you liked best? Are you curious of some obvious move that engines didn't seem to consider? Did you see a move in some variation and don't see why it can't be played at the root? Was there some very cool looking line that most engines liked but that nobody likes now, except for you?

It's time to force all these moves of your own into the root position, main lines and side lines, as a 1400 elo player I can tell you that very few chess knowledge is necessary to override the engines in specific lines of the variations, or even the main move, specially if you're trained your gut to tell you about some pawn push or threat that no engine considered but that turns out to be best. You would not be able to play the best continuation OTB but here finding a simple move may be enough to guide the engines.

And remember, in most positions there is no best move to find, as most of them are playable, so what matters is having a continuation in mind, and engines are only there to show you their variations and suggest goals with their evaluations, but chess isn't like those test positions you see in where there's a clear best move to be found while the rest suck, in real life there positions usually don't even appear on the board as they were seen from the distance and avoided.