Context: Sepsis is one of the leading causes of death in hospital settings. Timely administration of rational and effective antimicrobial therapy, as per hospital’s antibiotic policy, is one of the components of antimicrobial stewardship program. In absence of definitive pathogen identification and susceptibility pattern, initial antibiotic regimen is selected which is defined as empirical antibiotic therapy. Although an empirical antimicrobial policy is in place at our institute, it has not been evaluated since its inception for its effectiveness. Aims: To assess the adherence to the empirical antimicrobial policy for sepsis and to evaluate its effectiveness on clinical outcome of sepsis in adult critical care patients. Setting and study design: A prospective, cross-sectional study was conducted in adult non covid critical care units of Shree Krishna Hospital. Methodology: Following approval from Institutional Ethics Committee, prospective crosssectional study was conducted from 1st August-2021 to 31st July-2022 at non-covid intensive care units. The adherence to the empirical antibiotic policy was calculated as percentage of patients with sepsis in whom antimicrobial agent was started as per the policy. The effectiveness of the antimicrobial policy was assessed on the basis of the improvement in the clinical and laboratory parameters as well as Sequential organ failure assessment (SOFA) score of the patient over a period of five days. Statistical analysis: Microsoft Excel 2019, Version 2209 was used for data entry and data analysis. Proportions were calculated using descriptive analysis. Data was analysed using Chi-square calculation. Significance was considered at P-value <0.05. Results: The adherence to the antimicrobial policy in sepsis was 59.80%, (n = 61 out of 102) whereas adherence to initiation of antimicrobial agent within one hour of diagnosis of clinical sepsis was 96.07% (n =98 out of 102). The antimicrobial agent started as per antibiotic policy was susceptible in culture report in 55.31% (n = 26 out of 47) of patients. Effectiveness of empirical antimicrobial policy in patients with sepsis in adult critical care units based on improvement of SOFA score after five days of diagnosis of clinical sepsis was 51.06%, (n=24 out of 47). There was no significant correlation (P ≥0.05) found between age groups, gender, risk stratification categories, type of blood stream infections and type of organisms isolated, i.e., Gram negative and Gram positive with regards to effectiveness of empirical antimicrobial policy. There was no significant difference noted between improvement in SOFA score of the patients in whom antimicrobial agent was started as per policy and in whom the antimicrobial agent was not started as per policy (P = 0.72). Conclusion: The adherence to antimicrobial policy for sepsis was low and further studies to evaluate the reasons for low compliance need to be conducted. Although the effectiveness was not significantly different when the antimicrobial agent was started as per policy compared to when it was not started as per policy, we still recommend using antimicrobial agent as per policy in order to avoid non uniformity in prescriptions and development of antimicrobial resistance.