In this paper an approach for automatic detection of segments where a regression model significantly underperforms and for detecting segments with systematically under- or overestimated prediction is introduced. This segmentational approach is applicable to various expert systems including, but not limited to, those used for the mass appraisal. The proposed approach may be useful for various regression analysis applications, especially those with strong heteroscedasticity. It helps to reveal segments for which separate models or appraiser assistance are desirable. The segmentational approach has been applied to a mass appraisal model based on the Random Forest algorithm.