PERIT-S2S · Travel & hospitality

Rebookings during disruption — airport noise, urgency, and three changes in one turn.

Audio comes from airline desks, hotel front office, disruption lines. Scored on task success, with a working travel & hospitality operator as the control.

880
Clips in the target held-out set
68h
Audio, never published
8
Languages in the set
1
Working travel & hospitality operator as control

Leaderboard

Illustrative figures. Public runs begin when the held-out sets close.
ModelTask success (%)Median turn latencyBarge-in recovery
Human control · sector agentHuman control
93.0 ± 1.6
576ms95.9%
Frontier realtime · lab AFrontier
65.5 ± 1.8
819ms71.4%
Frontier realtime · lab BFrontier
57.3 ± 1.2
714ms56.8%
Cascaded stack (ASR → LLM → TTS)Specialist
51.6 ± 0.8
1514ms46%
Frontier multimodal · lab CFrontier
49.0 ± 1.3
828ms56.9%
Open weights · 8B speechOpen weights
29.7 ± 0.7
1140ms35.5%
Task families

What the set actually asks for.

  • Turn-taking & barge-in
  • Task completion
  • Instruction adherence
  • Disambiguation & repair
  • Escalation & refusal
  • Tone under pressure
Method & limits

How a run on this set is produced.

  1. 01Callers are scripted personas played by trained speakers, not synthesised — including the ones who talk over the agent.
  2. 02Every call is scored against the rubric by two reviewers; a third adjudicates splits.
  3. 038 calls per persona per model. Reported intervals are bootstrapped at 95%.
  4. 04Human control is a licensed agent from that sector handling the same personas blind.

Limits, stated up front: 8 languages only, single-channel audio, and no cross-sector transfer. The control is one operator per sector, so the human line carries its own error bar — it is a working standard, not a ceiling.