Step 1: Contrast pre-selection use of auxiliary information with post-selection use. Ratio and regression estimators are classic examples of post-selection use, where the auxiliary variable's values are used only after the sample data are already collected, to adjust the estimate.
Step 2: Pre-selection use means the auxiliary variable shapes the sample design itself, before any unit enters the sample. Forming strata is exactly this: the population is partitioned using known values of an auxiliary characteristic, and only then is random sampling carried out within each stratum.
Step 3: None of systematic sampling (needs only a list and a fixed interval $k$), cluster sampling (uses pre-existing natural groups, not necessarily an auxiliary study-related variable), or two-stage sampling (selects first-stage units then sub-units, generally without auxiliary stratification) requires this kind of pre-selection auxiliary input in the way stratification does.
Hence the technique using auxiliary information before sample selection is $\text{stratified random sampling}$.
\[\boxed{\text{Stratified random sampling}}\]