One of the approaches to conserving energy in WSN is to schedule some of the sensor nodes to sleep and estimate their readings from the ones of active nodes. This prediction relies on the modelling of the relationship between the sensor readings of the active and sleeping nodes. As this relationship can change over time, it is crucial to detect these changes to minimise loss in sensing fidelity. To address this problem, current approaches monitor the sensor readings of sleeping nodes at a regular interval, which can be wasteful of energy. In our protocol, BLESS (LessBattery), we propose an energy efficient monitoring technique based on the trust we have in the prediction model for future estimates of the sensor readings.