Digital transformation is still an ongoing process that changes how certain tasks are carried out in society. Not only is reshaping processes for the digital domain challenging, it has to be done in a way that includes everyone in social interaction, regardless of their perceptual, cognitive, or physical abilities. Designing processes and their user interfaces with digital participation in mind is crucial to avoiding a digital divide. Web interfaces still are perhaps the most important kind of user interface. Criteria have been defined to evaluate the accessibility of a service. However, many online services remain inaccessible to people with disabilities. Government websites have been required to be accessible for quite some time. Recent laws require companies to fulfill the same set of accessibility criteria. Therefore, testing user interfaces with respect to accessibility has become a mandatory part of website development. However, standard user interface testing is usually labor-intensive and cannot be fully automated. Thus, new approaches to user interface testing that focus on accessibility need to be researched. This paper proposes using Robotic Process Automation enhanced by Artificial Intelligence for accessibility testing. Although some automation tools have begun to incorporate artificial intelligence, a more fine-tuned approach to this kind of testing seems appropriate. Initial experiments were conducted on a limited set of test cases. The results suggest that automated web accessibility testing may be more effective with this approach.