A two-step numerical strategy based on a district digital twin is presented to efficiently deploy a limited number of depolluting panels in urban areas. In a diagnosis stage, a detailed pollutant concentration map is computed using CFD to identify critical highly polluted areas. Then, in a remediation stage, the optimal placement of depolluting panels as regards of urban airflow is determined to locally mitigate the air pollution in the exposed areas. For this purpose, a spatial sensitivity indicator calculated from an adjoint framework is proposed. The approach is applied to two real case studies: the full-scale laboratory district "Sense-City" under controlled conditions and a district area in Paris using realistic NOx traffic emission and wind conditions. In both case studies, it is shown that depolluting panels should be placed on a part of the sidewalks, the building facades and the roads adjacent to the sidewalks to reduce the high NOx concentration on some sidewalks and on first-floor building windows, thus preventing outdoor/indoor pollutant transfer. It is also proven that the proposed strategy is more efficient than a non-smart massive deployment of depolluting panels to improve the air quality in exposed areas. In addition to practical recommendations, this numerical strategy can provide a help-decision tool for city managers to design depolluting panels-based mitigation actions.