Leucine is the switch that turns muscle building on, so hitting your leucine threshold at every meal is what decides whether you grow.
Leucine really does start the response. Pooled across nine trials in older people it raised the muscle protein fractional synthetic rate substantially, a standardised mean difference of 1.08 with a confidence interval running from 0.50 to 1.67, and topping a small 6.25 g whey drink up to 5.0 g of total leucine made that drink match a 25 g whey drink for four and a half hours. What it does not do is finish the job. The same meta-analysis found no change in lean body mass, 0.18 with an interval from minus 0.18 to 0.54, or in leg lean mass. Three grams a day of free leucine across eight weeks of resistance training, in people already eating adequate protein, produced the same strength and the same quadriceps cross-sectional area as cornstarch.
Leucine is a switch, and the marketing stops there. Inside one meta-analysis the synthesis measure moved and the size measure did not, which is the distance between flipping a switch more often and having more to build with.