Applying principles of swarm intelligence to the control of autonomous systems in industry can advance our ability to manage complexity in prominent and high-cost sectors—such as transportation, logistics, and construction. In swarm robotics, the exclusive use of decentralized control relying on local communication and information provides the key advantage first of scalability, and second of robustness against failure points. These are directly useful in certain applied tasks that can be studied in laboratory environments, such as self-assembly and self-organized construction. In this article, we give a brief introduction to swarm robotics for a broad audience, with the intention of targeting future industrial applications. We then present a summary of four examples of our recently published research results with simple models. First, we present our approach to self-reconfiguration, which uses collective adjustment of swarm density in a dynamic setting. Second, we describe our robot experiments for self-organized material deployment in structured and semi-structured environments, applicable to braided composites. Third, we present our machine learning approach for self-assembly, motivated as a simple model developing foundational methods, which generates self-organizing robot behaviors to form emergent patterns. Fourth, we describe our experiments implementing a bioinspired model in a robot swarm, where we show self-healing of damage as the robots collectively locate a resource. Overall, the four examples we present concern robustness, scalability, and self-X features, which we propose as potentially relevant to future research in swarm robotics applied to industry sectors. We summarize these approaches as an introduction to our recent research, targeting the broad audience of this journal.