Asynchronous AI is becoming increasingly necessary due to the long response times of generative AI APIs, which can take tens of seconds or even minutes to respond, making synchronous technologies and paradigms unsuitable for these products. The traditional REST API model, designed for fast responses, is not well-suited for GenAI APIs, which require asynchronous solutions to handle their variable and often long response times. To address this issue, developers are exploring event-driven systems using callback events such as webhooks, streaming protocols like Server-Sent Events (SSE) or WebSocket, and adopting asynchronous AI patterns that better match the nature of GenAI products, offering improved developer experiences, scalability, and robustness.