Text this: Fostering effective hybrid human-LLM reasoning and decision making