A large language model would use prior chats and texts uploaded by the user as a dataset to train and generate a personalised version of the (LLM). Going a step further than predictive text, the LLM would use the previous context of the chat to generate an entire response. This would save huge amounts of time and allow for quick and personalised responses. This would be incredibly useful not only for everyday smartphone users but also for professionals who send dozens of messages daily or people that experience difficulty typing, whether that be people with reduced motor functions or individuals with other communications or accessibility issues.A