Scientific observation is a disciplinary-informed way of looking at the world that requires the coordination of domain knowledge and perceptual skills with specialized tools and techniques to systematically identify objects, organisms, specimens, or phenomena of interest. Identification is a particular form of skilled observational practice where the observer recognizes key features for classification. Learning to see like a scientist is crucial for many citizen science activities, yet participants often have limited access to training and resources needed to practice. We posit that observational fluency development in taxonomic identification activities for citizen science has two components: (1) increased accuracy in noticing relevant features, and (2) use of more disciplinary-based explanatory language; and self-efficacy is crucial while developing these competencies. This study explores the extent to which an interactive visual identification platform supports learners' observational fluency development in the context of aquatic macroinvertebrate identification in environmental education. We examined changes in water quality monitoring volunteers' accuracy, self-efficacy, and explanatory language use after participation in a training that integrated the online identification (ID) tool. Methods included surveys, focus groups, and an ID accuracy task before and after training. Our findings indicate that learners, especially novices, demonstrated increased self-efficacy and accuracy in specimen identifications after participating in a training utilizing the online tool. They also shifted their ID justifications, incorporating more scientific terminology and salient diagnostic characters, and fewer generic descriptive adjectives. Results suggest that trainings using dynamic visual tools can positively impact observational fluency development and the quality of data contributed by volunteers.