Background Agent Latency

Accept some processing latency by running AI agents in the background and only engaging with results when notified, rather than attempting real-time interaction.