The complex networks of today involve algorithms, humans, machines, and tools. As a consequence of their digital and sociotechnical nature, they increasingly exhibit autonomous and intelligent characteristics. This development puts prevailing assumptions about information systems at stake. In this paper, we specifically engage with a radical reframing of past conceptualizations of machine intelligence or agency. Based on our theorizing, we advance and substantiate a novel algorithmic agency perspective that centers around three key characteristics: algorithmic interpretation, data generativity, and input translation. These characteristics pave the way for a new wave of research in information systems. Our proposed research opportunities for data analytics of digital traces address conditions for, features of, and approaches to algorithmic agency.