Over the past decade, robotic process automation (RPA) emerged as a lightweight paradigm to automation in business enterprises, making automation more accessible to non-techie business users. In the industry, RPA vendors did not only provide out of the box RPA bots to automate manual tasks on legacy software, they also provided users a recorder to create their own bots for specialized tasks. Unfortunately, if the recorders do not create generalizable models, users may face a bot sprawl and governance problem. In this work, we survey how the fields of artificial intelligence, and robotics specifically, approached the generalizability problem. Then, we discuss how RPAs can leverage those advancements to create generalizable bots from user's instructions or demonstrations.