The mainstream view in contemporary analytic philosophy is that perception is primarily in the business of representing the mind-independent world as it is. This work explores an alternative conception: that the goal of perception is to guide successful action and that perceptions do not need to track mind-independent properties to play this action-guiding role. I focus on two types of perception: color perception and pain perception. I start with the former and advocate a pragmatist, empirically-guided approach which begins by inquiring into the function of color vision. After arguing that none of the extant philosophical views of color are satisfactory, I answer the function question by focusing on systematic color perceptual phenomena investigated by psychophysicists. I argue that the human color visual system is an enhancement system: that is, its job is to help us better discriminate, track, and recognize meaningful objects, properties, and relations. I then build on this idea using the notion of ‘competence-embeddedness.’ I propose that color vision is embedded in a network of competences: the aim of color vision is to help organisms manifest these competences, and color experiences are correct when they result from competence-enhancing processing. The framework is explanatorily robust. For example, it allows me to conceptualize many textbook color illusions as special cases of successful color perception where the demands of the relevant competences clash. Finally, I use the notion of ‘competence-embeddedness’ to develop a new account of pain. I argue that the pain system is not a bodily disturbance detector, but a sophisticated, context-responsive security system whose primary goal is to help organisms manifest important behavioral and cognitive competences.