Why not just use a call center for this? Someone pulls up to the speaker, and someone in another part of the world takes their order and puts it into the restaurant’s system. Surely that would be easier and have a greater chance of success than voice gen?