Skeptical invariantists maintain that the expression “knows” invariably expresses an epistemically extremely demanding relation. This leads to an immediate challenge. The knowledge relation will hardly if ever be satisfied. Consequently, we can rarely if ever apply “knows” truly. The present paper assesses a prominent strategy for skeptical invariantists to respond to this challenge, which appeals to loose talk. Based on recent developments in the theory of loose talk, I argue that such appeals to loose talk fail. I go on to present a closely related response strategy, which appeals to conversational exculpature, a phenomenon recently studied by Hoek. A promising account of knowledge ascriptions results, which combines the virtues of invariantism and elaborate versions of contextualism, positing a unique knowledge relation while still making precise predictions on when knowledge ascriptions ordinarily count as appropriate.