Alexa, Siri, and other #voice assistants rely on turn-based “verbal ping-pong,” waiting for users to finish speaking before responding. Humans communicate continuously, updating understanding in real time. Incremental agents mimic this by processing speech word-by-word, enabling more natural interaction. In this talk, Casey Kennington explores the need and requirements for building real-time processing into agents.

🎬 https://youtu.be/I26KBfY3Vx8

More in: https://www.w3.org/2025/10/smartagents-workshop/report.html
#SmartVoiceAgents

Handling hallucinations in voice agents can be even more challenging than in text-based #chatbots. In this talk, Ulrike Stiefelhagen shows that industry (Workers Daily Summary) and healthcare (Patient Chat) use cases reveal how evolving requirements in #LLM systems can actually create new opportunities for #voice, even in traditionally difficult contexts like #privacy and noisy environments.

🎬 https://youtu.be/nGf2mvRV2QI

More in: https://www.w3.org/2025/10/smartagents-workshop/report.html
#SmartVoiceAgents

Voice AI moved from #VoiceXML to LLM-driven agents, creating powerful but fragmented platforms. In this talk, RJ Burnham explores whether it’s time for new #voice #AI standards that restore interoperability without stifling innovation.

🎬 https://youtu.be/wSbdFGzZZhs

More in: https://www.w3.org/2025/10/smartagents-workshop/report.html
#SmartVoiceAgents

Pronunciation in web content suffers from the lack of a standardized way to specify Synthetic Speech Markup Language (SSML) in #HTML. A standard would benefit assistive technologies, voice agents, and #AI systems learning from web content. In this talk, Sarah Wood showcases an EdTech standards community solution and explores ways to support consistent, accessible pronunciation in web #voice applications.

🎬 https://youtu.be/Bc04fXrR1U4

More in: https://www.w3.org/2025/10/smartagents-workshop/report.html
#SmartVoiceAgents

In this talk, Raj Tumuluri explores the gap between functional performance and emotional resonance in modern #AI, outlining principles for building empathetic, trustworthy systems. Topics include tone consistency, handling ambiguity, and design choices that convey reliability, concluding with a practical framework.

🎬 https://youtu.be/AJMoRjPXn-g

More in: https://www.w3.org/2025/10/smartagents-workshop/report.html
#SmartVoiceAgents

In this talk, Zohar Gan explores improving #accessibility of 3D web content through semantic metadata and #voice interaction, with a demo and a proposal to standardize a semantic 3D metadata schema, its embedding in media and web spatial voice.

🎬 https://youtu.be/Cq2Is-uIuNU

More in: https://www.w3.org/2025/10/smartagents-workshop/report.html
#SmartVoiceAgents

Screen readers enable non-visual web access but lack conversational context. In this talk, Brian Vuong demonstrates how #AI #voice agents can bridge the gap between visual interfaces and blind/low-vision users. He identifies the specific #WebStandards gaps that currently hinder the seamless integration of such third-party agents into the DOM.

🎬 https://youtu.be/qnnYe6D9Cgc

More in: https://www.w3.org/2025/10/smartagents-workshop/report.html
#SmartVoiceAgents

The growth of voice agents across platforms is hindered by usability, security, and interoperability challenges. In this talk, Patricia Lee presents a data-driven framework for Web standardization, focusing on stakeholder needs, trust and compliance, and measurable requirements to enable scalable, user-friendly #voice #AI adoption.

🎬 https://youtu.be/kT_Z5NWyi5Q

More in: https://www.w3.org/2025/10/smartagents-workshop/report.html
#SmartVoiceAgents

The report from the W3C Workshop on Smart Voice Agents, held online in February 2026 is now out.

Participant RJ Burnham, summarized a central issue:
"Proprietary voice AI platforms can move quickly, but the result is fragmentation and lock-in. The key question is whether we can restore portability and interoperability without slowing innovation."

Read more including eight cross-cutting issues that now define the practical standards agenda at:
https://www.w3.org/news/2026/w3c-workshop-report-smart-voice-agents/
#SmartVoiceAgents

In this talk, Kristiina Jokinen explores designing trustworthy #GenAI based applications, focusing on grounding as a key principle for spoken interaction. She reviews challenges, opportunities, and lessons learnt in creating accountable reasoning and fluent natural dialogue between smart AI agents and users, to support long-term interactions and #trust for responsible and safe #AI agents.

🎬 https://youtu.be/8Pmxmn7gCsA

More in: https://www.w3.org/2025/10/smartagents-workshop/report.html
#SmartVoiceAgents