Experiments with 40 volunteers show that robots using Visual Language Models can adjust behavior based on facial expressions and context. This approach improves how humans perceive robotic collaborators during shared tasks. The study demonstrates a shift toward context-aware interaction. Practitioners can now integrate environmental cues to make robotic responses feel more natural.