What is latency hiding in an Agent's user experience?

• A. Reducing the number of tool calls to avoid delays.
• B. Preloading all data before starting any interaction.
• C. Batching tool results into a single response after completion.
• D. Making the system feel responsive by using asynchronous calls, streaming outputs, or progress updates.

Answer: (D)

Latency hiding means making the user experience feel fast by masking delays during interaction. The key idea is to keep the system responsive in perception, not just by making things faster behind the scenes, but by providing feedback and progress while work is still underway. Techniques like asynchronous calls let the UI stay interactive, streaming outputs reveal results as they’re produced, and progress updates or typing indicators keep the user informed. This approach creates a sense of momentum and reduces the impression of waiting, which is often more important for a smooth experience than shaving off every millisecond of actual processing time. Other strategies—such as reducing the number of tool calls, preloading all data before interaction, or delivering results in a single batch after everything finishes—address latency differently and don’t inherently hide the waiting in real time.