Using chemical dissolution and heat melting to remove material, Electrochemical Discharge Drilling (ECDD) is a reliable method for processing non-conductive materials. Even if it works well, it might be difficult to maintain an efficient machining gap during hole drilling owing to contact of tool with the material. This can result in poor machined surface geometric features such deteriorating hole shape, poor surface quality and increased thermal cracking. The objective of this study is to improve geometric features and machining performance by integrating a cognitive tool feed approach into ECDD. The cognitive tool feed enhances the overall geometrical aspects of the micro-holes by utilising the principle of tool retraction once it makes contact with the material. To confirm its effectiveness, the acquired features of the drilled holes are examined and compared with and without cognitive tool feed system. The impact of several control variables on response characteristics is also examined, encompassing Circularity Error (CE), Surface Roughness (SR), Hole Overcut (HOC), Material Removal Rate (MRR), Thermal Cracks and Number of Tool Contacts (NTC).