A voice agent can have a strong model and still fail because it talks over someone, leaves a dead pause, or hands a caller to a human with no useful context. These talks follow that failure back through the system. They begin with conversation design: what the agent is trying to accomplish, how it knows a turn is over, and when an interruption should stop it. The route then moves through architecture and media transport, where every model call and network hop spends part of a tight timing budget. Smaller models, state machines, colocated services, and faster speech inference buy some of that time back. The final talk puts the pieces inside a support operation, where monitoring, escalation, and context handoff decide whether a polished demo can answer a real phone line.
Why first: Seagull shows how transcription, streaming, drift and subjective evaluation make ordinary agent problems harder. Caryotakis follows by deciding which of those problems matter to the customer on the other end of a call.
Why second: Seagull inventories the ways a voice agent can fail. Caryotakis makes a product choice among them: finish the conversation reliably before chasing a more realistic voice, then read the script aloud to find the awkward parts.
Why here: Caryotakis asks teams to design the conversation as spoken behavior. Shapland explains the mechanism that most often ruins that behavior: a serial pipeline that mistakes a short silence for the end of a turn.
Why here: Shapland explains why endpoint prediction is hard. Agrawal and Wirjo compare three ways to make that decision in a working pipeline, then show why a cough, an acknowledgement and a correction should not trigger the same response.
Why here: Agrawal and Wirjo settle how the user gets the floor. Sherbakov and Kotha widen the choice to the whole agent, weighing chained and speech-to-speech systems and delegating difficult work away from the frontline voice model.
Why here: Sherbakov and Kotha choose what runs inside the agent. DuBois and Kramer turn to the path between user and model, where ordered delivery, packet loss and jitter can make a fast architecture feel slow.
Why here: DuBois and Kramer fix the media path from the user's device. Bhargava budgets the delay that remains across speech recognition, the language model and speech generation, then shows what colocation and a thinker-talker split can save.
Why here: Bhargava keeps a small model on the conversational path and sends harder work elsewhere. Allou and Bahidika make the bargain concrete by moving lesson state into code so the speaking model can answer in time.
Why here: Allou and Bahidika cut the model's thinking time. Kiely finds the next constraint in speech serving, where first-byte latency and concurrent streams matter more than raw token speed, and careless client code can give the gain back.
Why last: Kiely gets the listening, thinking and speaking machinery into its latency budget. Bar puts it on a real support line, where short answers, filler speech, monitoring, escalation and a useful handoff decide whether callers keep using it.