video1.0<iframe src="https://www.loom.com/embed/a1fad8b0be694a529530f60ac1569ac5" frameborder="0" width="1720" height="1290" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>12901720Loomhttps://www.loom.com12901720https://cdn.loom.com/os/loom-media-content/loom/346ef557-93bf-4857-a6e5-2c240dea7494/29cca35d-ff33-4522-a07f-236f315ec9ca/sessions-thumbnails-a1fad8b0be694a529530f60ac1569ac5-d159b941f4888d3a.gif/binary?authToken=eyJraWQiOiJtaWNyb3MtZ3JvdXAvdGRwLW9zLWFyY2hldHlwZS9yYjVnOGJkdXJoZzQ4cDFmIiwiYWxnIjoiUlMyNTYifQ.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.Mr266OTgzcYCye8ISPhy7YSnWYQJBLZX9etEH2gDcpzvgqCtInHX6A30gDct7RluIZFDMBEissJwDm7aVJCDhR7hoxyoZNVXTzroVHNn0d_lqkTbyBG63MeHkRMofr5fGtgVLxi3UNsO1AnAyXdcDWdcJ9V9Zk58b44N4nMgn8loOSz_ZeuojWXWdlRg0ddUfe0mRviViOI2Gl6rwPuGs3_TOw_Q3bE-o6kHie7GfTFS3mC0JHYqzbBD0ripc7O0SAAncGqS3q6-Tqg_QEPlc5Iw8DCr3zZJ0xDHTh5uRAplv0iBtbTC7PJm2pJ5WdUSgyCi-CxsQXgWO5Q1AXMvIQ&Expires=1791643061&Signature=FmPmlyW3bydbcGepfBtjjBDpt9rjbC00QHWipgXqYbDbZenjKVk8x13x%7Et4N6AfBPuNHo%7E5dKTgc7CprItQGAQVuMwLGHE0V3bqrie4GnkTA%7E5o%7EwmcU1ZluFjSVk6oiLi8UzBW7OiHymqjELRnl0ShtAjY0qPWpVBCILgGND2IbZ4GsIPa3RNwBtfwpTIm0-3fc8U%7EfLPG9VEWybc0sdU3u35AWVqFaT9bZ3RtaeeI3LS%7E83IaAg0nngKmpaXTu6HUW3ssbOhLCXnhHewpg1o%7E9a20UlQxPZRhqT06sloP%7Evwj-ZQkPMmRBCsYI6PnS1kN4pp%7EIrEIqwVTL-QWSaw__&Key-Pair-Id=KQOSYIR44AIC0228.288This Loom explains how to configure the conversational experience and call controls for a voice AI agent, including knowledge base readiness and idle behavior. It notes that once a knowledge base reaches ready status, it can be attached so the agent answers queries from it. For conversational experience, it covers setting background noise with preview and volume controls, adding filler words, and balancing “eagerness to speak” for faster responses with token usage tradeoffs. It also describes idle conversation settings: after 15 seconds of no user speech the agent performs a pulse check, and if there is still no response it ends the conversation 20 seconds later, with an optional max call duration that provides a warning 15 seconds before ending or transferring to a human agent.