Imagine a robot assembling furniture and, running into a tricky step, projects instructions directly onto the parts you need to handle! That’s the promise of SiSCo, a system that combines the power of Large Language Models (LLMs) with mixed reality to turn robots into expert communicators. Traditionally, robots have relied on pre-programmed cues or rigid visual languages. SiSCo revolutionizes this by using LLMs to generate instructions on-the-fly, adapting to the specific context of the task. Researchers tested SiSCo in a human-robot teaming scenario, where a robot and a human worked together to assemble structures. The results were impressive. Using projected augmented reality signals, teams were able to solve problems 73% faster, with an 18% increase in success rate, compared to using basic language instructions. Not only were the teams more efficient, but they also found the visual cues easier to follow. SiSCo translated complex instructions into clear, projected visuals, reducing the mental effort required and improving overall comprehension. This fusion of mixed reality and LLMs shows how robots can become intuitive partners, understanding our needs and guiding us through collaborative tasks in a way we’ve never seen before. The open-source nature of SiSCo promises to unlock further development and push the boundaries of human-robot communication.