A mechanistic understanding of electro- and photocatalytic CO2 reduction is crucial to develop strategies to overcome catalytic bottlenecks. In this regard, for a new CO2-to-CO reduction cobalt aminopyridine catalyst, a detailed experimental and theoretical mechanistic study is herein presented toward the identification of bottlenecks and potential strategies to alleviate them. The combination of electrochemistry and in situ spectroelectrochemistry together with spectroscopic techniques led us to identify elusive key electrocatalytic intermediates derived from complex [LN4Co(OTf)2] (1) (LN4 = 1-[2-pyridylmethyl]-4,7-dimethyl-1,4,7-triazacyclononane) such as a highly reactive cobalt(I) (1(I)) and a cobalt(I) carbonyl (1(I)-CO) species. The combination of spectroelectrochemical studies under CO2, 13CO2, and CO with DFT disclosed that 1(I) reacts with CO2 to form the pivotal 1(I)-CO intermediate at the 1(II/I) redox potential. However, at this reduction potential, the formation of 1(I)-CO restricts the electrocatalysis due to the endergonicity of the CO release step. In agreement with the experimentally observed CO2-to-CO electrocatalysis at the CoI/0 redox potential, computational studies suggested that the electrocatalytic cycle involves striking metal carbonyls. In contrast, under photochemical conditions, the catalysis smoothly proceeds at the 1(II/I) redox potential. Under the latter conditions, it is proposed that the electron transfer to form 1(I)-CO from 1(II)-CO is under diffusion control. Then, the CO release from 1(II)-CO is kinetically favored, facilitating the catalysis. Finally, we have found that visible-light irradiation has a positive impact under electrocatalytic conditions. We envision that light irradiation can serve as an effective strategy to circumvent the CO poisoning and improve the performance of CO2 reduction molecular catalysts.