How do you handle failure states and API timeouts when communicating with external AI/ML services in a Develop Health system?
