The operational safety assessment of structures and mechanical components containing crack-like flaws remains among the most challenging aspects of engineering practice. Fitness-for-service standards, such as BS 7910 and API 579/ASME FFS-1, incorporate state-of-the-art methodologies for the structural integrity assessment of flawed structures, with the Failure Assessment Diagram (FAD) approach being the most widely adopted for the evaluation of individual cracks. Nevertheless, in many cases, multiple cracks are observed in close proximity, wherein the stress fields associated with each crack interact, potentially resulting in an amplification of the driving forces for crack propagation. The methodologies outlined in fitness-for-service standards recommend recharacterizing such defects as a single flaw – referred to as a bounding flaw – and subsequently conducting the assessment of this idealized defect using the FAD methodology. Recent findings, however, have indicated that this approach may be excessively conservative, thereby leading to the premature retirement of structures or components. Accordingly, the present study aims to compare the conventional bounding flaw approach with a more refined analysis of surface cracks under interaction regimes. The influence of geometric parameters, such as aspect ratio, depth ratio, and coplanar distance, was evaluated both in the construction of FADs and in the assessment of the effectiveness of the bounding flaw methodology. To this end, a series of finite element simulations was conducted. The analysis of the failure assessment diagrams revealed a reduction in the area beneath the curve – indicative of safe operating conditions – as the flaws approached one another, which was correlated with an increase in the plastic component of the J-integral induced by their interaction. However, as the depth ratio of one of the flaws increased, the influence of the coplanar distance was observed to diminish, as the relative amplification decreased in comparison to the other flaw. Ultimately, the bounding flaw methodology was deemed effective in ensuring operational safety, as it consistently exhibited greater severity than the highest amplification levels observed among the interacting flaw pairs. Nevertheless, conservative – and, in certain instances, overly conservative – scenarios were identified. Furthermore, it was observed that this excessive conservatism could be mitigated by considering the flaw with the greater depth ratio for the estimation of the reference stress and, consequently, for the determination of the load ratio coordinate.