It would be easy enough to become preoccupied in this position with the pinned Black bishop, but White has no way to exploit it. If he throws a bishop at it, as with Bc3, the trouble is not merely that Black can answer with Rd6; it's that Black then has a check in Rb1+, White having opened a line to his king when he moved his bishop. White then must interpose his bishop at e1, as all of his king’s flight squares are under attack. With White’s bishop now pinned and Black to move, he hits the immobilized piece with a pawn: f3-f2. Black wins.
The key to the position for White is to step back from the pin and look for other patterns. Black’s bishop and rook are aligned on the same diagonal, and this is grounds to consider possible pins or skewers. White has a bishop that can attack on the dark-squared diagonal with Bf2, but this runs into the same trouble just described (Black’s reply of Rb1+—look for your opponent’s next check, not just your own). It would be different if White’s move gave check. Might this be arranged? Yes, with the substitution 1. RxB+, KxR. Another route to the same conclusion is to just consider any checks White can give; here there is only one—RxB. Black is forced to reply KxR, and with his king moved you look at the new lines it occupies and see that the kernel of a skewer now has been created. Bf2+ attacks Black’s king, requiring him to spend a move taking it out of the way. Then BxR takes Black’s rook before it has time to give check.