Agricultural soils are increasingly recognised as major sinks of microplastics (MPs), yet evidence from Latin America—and particularly from Colombian Andean cropping systems—remains scarce, and most regional laboratories must adapt international protocols to limited analytical infrastructure. This short communication reports detection of MPs in agricultural soils of Cachipay (Cundinamarca, Colombia), a temperate-humid Andean municipality, and evaluates a feasible, low-resource extraction and identification workflow. Composite topsoil from a common-bean (Phaseolus vulgaris) plot, divided into two field sectors (East and West), was processed by dry sieving, density separation, and vacuum filtration, and the recovered particles were examined by attenuated total reflectance Fourier-transform infrared spectroscopy (ATR-FTIR). Two density media were compared under identical conditions: saturated sodium chloride (NaCl, ~1.2 g cm⁻³) and sodium iodide (NaI, ~2.1 g cm⁻³), with distilled water as a procedural blank. MP-compatible particles were detected in both sectors. NaI recovered particles in all replicates of both sectors, whereas NaCl yielded a positive signal only in the East sector; distilled-water blanks remained negative, supporting the specificity of the procedure. Diagnostic absorption bands were consistent with polyethylene, polypropylene, polystyrene, and polyethylene terephthalate (predominantly identified within the 600–800 cm⁻¹ and 1700–1800 cm⁻¹ wavenumber ranges) with additional amine-associated signals near 3400 cm⁻¹ tentatively attributed to a polyethyleneimine-type material. Because the workflow was qualitative, the study confirms presence or absence only and does not quantify particle abundance, size, or mass. These results provide baseline evidence of MP occurrence in an under-studied Andean agroecosystem and demonstrate that a denser separation medium (NaI) markedly improves recovery and the reliability of ATR-FTIR assignment in fine-textured, organic-rich soils, offering a practical template for resource-limited monitoring programmes.