Full AWS Practitioner Certification Question

What is the default way to consume an endpoint after deploying a real-time inference model on an endpoint?