Some scientists who study animals have emphasized the need to focus on the “point of view” of the animals they are studying. This methodological shift has led to animals being credited with much more agency than is warranted. However, as critics suggest, on the one hand, the “perspective” of another being rests mostly upon “sympathetic projection,” and may be difficult to apply to unfamiliar beings, such as bees or even flowers. On the other hand, the very notion of agency still conveys its classic understanding as intentional, rational, and premeditated, and is still embedded in humanist and Christian conceptions of human exceptionalism. This paper seeks, in the first part, to investigate the practical link between these two notions and the problems they raise. In the second part, following the work of two historians of science who have revisited Darwin's studies of orchids and their pollinators, it will observe a shift in the meaning of the concept of agency. Indeed, creatures may appear as “secret agents” as long as we adopt a conventional definition of agency based on subjective experience and autonomous intention. However, when reframed in the terms of “agencement” — an assemblage that produces “agentivity” — agency seems to be much more extensively shared in the living world. We will then explore some of the concrete situations in which these agencements are manifested, and through which creatures of different species become, one for another and one with another, companion-agents.