Ants are one of the most widespread groups of insects, but the mechanisms allowing them to coexist despite having similar nutritional needs remain unclear. In this study, we investigated the structure and competitive dynamics of ant assemblages in pineapple agrosystems of La Réunion Island. Using an imagery approach that combined in-field image captures and a computer vision algorithm based on convolutional neural networks, we analyzed ant dominance hierarchies, foraging and recruitment behaviors, niche partitioning between day and night, as well as the impact of environmental variables. This approach provided highly detailed data on the strategies of individual species. Additionally, to better understand the resource exploitation strategies employed by ant species, we then tested the discovery-dominance trade-off. Our results revealed that Pheidole megacephala and Solenopsis geminata were the two dominant species, displaying high recruitment investment and frequent bait monopolization, particularly at night. In contrast, Brachymyrmex cordemoyi used a scout-based foraging strategy that allowed it to rapidly locate baits and maintain moderate dominance levels, during both day and night periods. With lower abilities to discover and control baits, subordinate species had to rely on alternative foraging strategies to persist. We found a significant positive correlation between the discovery and dominance abilities supporting a discovery–defense strategy rather than a discovery-dominance trade-off. This finding implies that resource monopolization in simplified ecosystems may benefit species optimizing both rapid resource location and competitive defense. Environmental factors, including plant species richness, plant cover, and mulch cover, influenced species abundance and occurrence, even though dominant species appeared to remain unaffected by habitat variations. The impacts of environmental factors highlight the role of habitat complexity in mediating competitive inte