Structures normally experience a decline in performance over time due to material deterioration and external conditions. Ensuring structural integrity is critical for their longevity and safety, with concrete cracking being an indicator of damage. Manual inspection methods for detecting cracks are time-consuming and subjective, which has led to a decline in their effectiveness as the number of buildings increases. This study aims to address these issues by developing a more efficient method for detecting and documenting concrete cracks on building surfaces using deep learning algorithms. A database of 40,000 images was used to train and test various convolutional neural network (CNN) models. They are evaluated using quality assessment metrics derived from the confusion matrix. The most suitable model was then optimized and used to develop a graphical user interface (GUI) for rapid classification of new observations. This study was structured into four sections, starting with an introduction and then a brief presentation of the image database and the deep learning models. Their predictive capabilities were evaluated and compared to select a suitable model to optimize and create a GUI. The study concluded and discussed the results, demonstrating the potential of using a deep learning model as a tool for automated concrete crack detection that is suitable for integration into maintenance workflows.