•We studied whether fNIRS could differentiate between different levels of difficulty in everyday representations of reading and writing tasks of information workers.•We used personalised tasks to each participant, with simulated interruptions to increase ecological validity of task within a laboratory study.•Subjective self-reports showed differences for both reading and writing tasks, where the hard dual-task level was significantly different from easy and medium levels.•fNIRS over the prefrontal cortex was able to discern the same difference for reading, but not for writing.•These results could be explained by the neural areas in which mental workload was presented for both tasks.•fNIRS detected some differences in brain activity which corresponded to the interruptions, and the findings can be understood in terms of spare capacity models. The motivation behind using physiological measures to estimate cognitive activity is typically to build technology that can help people to understand themselves and their work, or indeed for systems to do so and adapt. While functional Near Infrared Spectroscopy (fNIRS) has been shown to reliably reflect manipulations of mental workload in different work tasks, we still need to establish whether fNIRS can differentiate variety within common office-like tasks in order to broaden our understanding of the factors involved in tracking them in real working conditions. 20 healthy participants (8 females, 12 males), whose work included office-like tasks, took part in a user study that investigated a) the sensitivity of fNIRS for measuring mental workload variations in representations of everyday reading and writing tasks, and b) how representations of natural interruptions are reflected in the data. Results supported fNIRS measuring PFC activation in differentiating between workload levels for reading tasks but not writing tasks in terms of increased oxygenated haemoglobin (O2Hb) and decreased deoxygenated haemoglobin (HHb), for harder conditions compared to easier conditions. There was considerable support for fNIRS in detecting changes in workload levels due to interruptions. Variations in workload levels during the interruptions could be understood in relation to spare capacity models. These findings may guide future work into sustained monitoring of cognitive activity in real-world settings.