DuplexWorld: Can voice agents help you get through the day?

Source

• Source name: arXiv
• Source kind: paper
• Authors / organization: Aryan Vijay Bhosale, Harshit Rajgarhia, Akhil Pothanapalli, Asif Shaik, Abhishek Mukherji, Dinesh Manocha
• Venue: arXiv: cs.SD
• Published: 2026-08-11
• Collected: 2026-08-12
• URL: http://arxiv.org/abs/2608.10716v1
• PDF: https://arxiv.org/pdf/2608.10716v1
• DOI: UNKNOWN
• Registry group: academic_apis_and_preprints
• Evidence role: canonical_academic_index
• Language: en
• Country / region: UNKNOWN

Plain-Language Brief

A research paper titled 'DuplexWorld' discusses the use of speech-to-speech (S2S) voice agents in various sectors like banking, insurance, travel, healthcare, logistics, and pathfinding. The study evaluates the performance of these agents based on conversational and analytical capabilities across 156 scenarios. It concludes that while the best voice agents show room for improvement in all areas, they fail to adequately handle real-world conversations and tasks beyond database manipulation.

What the Document Says

• The study focuses on speech-to-speech (S2S) voice agents used in customer care and daily life.
• Existing benchmarks for evaluating voice agents are criticized for not accounting for the diversity of conversational dialogue in mundane activities and for not testing agent's ability to assist with tasks beyond database manipulation.
• The study introduces six worlds: banking, insurance, travel, healthcare and logistics, and Pathfinding, where voice agents are especially useful.
• Agents are evaluated on eleven different types of conversations across 156 scenarios (350+ hours of conversation).
• The evaluation measures agentic, conversational, and speech-naturalness metrics.
• Results show that the best voice agents require improvement in all areas: Pass@1 (0.490), turn-taking (0.653), and DNSMOS (3.378).

Key Concepts Mentioned

• speech-to-speech voice agents
• customer care
• daily companions
• benchmarks
• DuplexWorld
• evaluation metrics
• Pass@1
• turn-taking
• DNSMOS

Claims Made by the Document

• Existing benchmarks fail to holistically evaluate voice agents.
• Voice agents need improvement in conversational and analytical capabilities.

Evidence Used

• Failure of existing benchmarks to handle diversity of conversational dialogue and tasks beyond database manipulation.
• Results showing substantial room for improvement in all areas for the best voice agents.

Methods / Approach

UNKNOWN

Limitations Stated by the Authors

UNKNOWN

Limitations Noted by the Agent

Agent interpretation — not from the source.

UNKNOWN

Open Questions

UNKNOWN

Notes for Future Q&A

May include interpretation. Labeled where it does.

UNKNOWN

Raw Abstract / Excerpt

Speech-to-speech (S2S) voice agents are increasingly being incorporated into enterprise for customer care and as daily companions for consumers owing to the ease of the conversational modality over text. However, existing benchmarks fail to holistically evaluate voice agents along axes that really matter and are shaped as tests of agentic tool calling against a database. We believe they fail to adequately account for the diversity of conversational dialogue that mundane activities introduce and further, never test how faithfully an agent can assist on tasks that move beyond database manipulation. To tackle this DuplexWorld introduces six worlds where voice agents are especially useful: banking, insurance, travel, healthcare and logistics, and Pathfinding. Agents are evaluated on eleven different types of conversations across 156 scenarios (350+ hours of conversation), each testing conversational and analytical capability to varying degrees. Through extensive evaluation comprising agentic, conversational and speech-naturalness metrics, we show that even the best voice agents leave substantial room for improvement on all 3 axes (Pass@1: 0.490, turn-taking: 0.653, DNSMOS: 3.378). We perform extensive analysis on agentic v conversational performance, world- and conversation type-wise performance, failure modes exploring the explore v exploit lens for Pathfinding conversations and voice agent reliability over all six worlds.