The construction boom underway in our seas is contributing to a loss of biodiversity. Built structures often have flat, featureless surfaces that provide little protection from predators and environmental stressors, and a relatively homogenous environment for colonisation. Increasingly habitat complexity is being eco-engineered into built structures to bolster biodiversity. There are, however, few tests of whether habitat complexity-biodiversity relationships are universally positive across environmental gradients, particularly on built structures. We compared the effects of habitat complexity on seawall biodiversity across two gradients: intertidal elevation, and the estuarine gradient of salinity. Across the intertidal gradient, we found effects of complexity that generally increased with abiotic stress. At high and mid intertidal elevations, where complexity strongly mitigated temperature extremes, we found positive effects of complexity on biodiversity and cover. However, at low intertidal elevations, where temperature maxima were less extreme and complexity had less effect on microclimate, we found neutral effects of complexity. Across the estuarine gradient, effects of complexity generally decreased with distance from the estuarine mouth, perhaps reflecting the diminishing species pool of available colonists on which complexity could act. Our results reinforce the need to consider key stressors to organisms and the niches of colonsists when designing eco-engineering interventions.