The widespread use of corticosteroids as molecules with high therapeutic potential has shown a gradual increase in recent years, but there is a need to progressively have molecules with high therapeutic effects but an increasingly low side effect profile. The chemical approach to the synthesis of new corticosteroids is often time-consuming, expensive, impacts the environment, difficult, and requires multiple manufacturing steps in contrast to new green approaches. In recent decades, microbial biotransformation has introduced an alternative to chemical synthesis of corticosteroids; however, although biotransformation seems to be a good alternative approach to traditional synthetic routes, the products often require difficult downstream purification processes. The purpose of this work was to identify an alternative process to liquid-liquid extraction for the isolation of two new steroids, 1,9β,17,21-tetrahydoxy-4-methyl-19-nor-9β-pregna-1,3,5(10)-trien-11,20-dione (SCA) and 1,9β,17,20β,21-pentahydoxy-4-methyl-19-nor-9β-pregna-1,3,5(10)-trien-11-one (SCB), obtained by biotransformation of cortisone with Rhodococus rhodnii. Cloud point extraction was chosen and applied for the first time as a “green” downstream method, as it is characterized by a profile of low cost, short extraction time, aqueous environment, moderate temperature, as well as performed with absolutely non-impactful and non-toxic extraction agents. Therefore, this study allowed the evaluation of the main factors affecting cloud point extraction, such as the concentration of polyethylene glycol tert-octylphenyl ether (TX-114), chosen as the extraction agent, of the electrolytes, such as NaCl, of the temperature and equilibration time at which the extraction is performed. Evaluation of extraction efficiencies was performed by RP-HPLC with values of 96% for SCA and 86% for SCB in real biotransformation samples. The analytical performances of the method were evaluated in terms of limits of detection (0.022 g L-1 for SCA and 0.016 g L-1 for SCB), limits of quantification (0.069 g L-1 for SCA and 0.050 g L-1 for SCB), linearity (0.072–3.82 g L−1). The relative standard deviations were calculated for the repeatability (0.54–0.91% for SCA and 0.70–3.80% for SCB) and stability (0.62–0.80% for SCA and 2.21–2.54% for SCB) of the proposed method. The accuracy was evaluated through recovery studies in relation to 50%, 100%, 150% excess analytes, allowing to obtain 99.69–99.77% recovery of SCA and 99.24–99.51% for SCB with RSD% less than 5%.