Credit scoring model is a key element of the automated loan approval system. The developed model is required to be validated under the internal rating-based approach of the BASEL framework. There are several techniques available for validating statistical models. These include KS statistics, C statistics, and HL statistics. When the bank cannot build a statistical model either because it does not have high quality data or it decides to rely on the experts’ judgement for certain types of loans, the validation of their credit rating system can be problematic. This paper proposes the use of AHP technique to develop credit scoring model based on experts’ judgement. Since the consistency ratio or the geometric consistency index under AHP technique can test only the expert’s consistency in their priority settings at the time the model is built, this paper proposes the actual decision validation of the AHP based on the concept of KS statistics. The lower and cut-off scores can then be computed using the bank’s acceptable inconsistency rate.