ElevenLabs vs PolyAI

Sequel Launches AI Agents Combining Website and Webinar Data

Sequel, a company that transforms webinars and video into a go-to-market engine, has announced the general availability of Sequel Agents, AI agents for webinar and video marketing powered by first-party audience context. These agents help marketing teams plan programs, create content, understand their audience, and act on engagement, extending the value of each webinar beyond the live broadcast. The agents use combined first-party website and video engagement data to publish recaps, adapt website experiences, and draft follow-ups. Customers can start with three types of agents (Build, Scale, Convert) or build custom agents. All agent workflows require human approval before sending emails, making changes, or pushing leads to CRM. ElevenLabs generated $14 million in attributed pipeline using the technology, and Unit21 automated 80% of post-webinar work.

• Sequel announced general availability of Sequel Agents for webinar and video marketing.
• Sequel Agents use first-party website and video engagement data to plan programs and create content.
• Customers can start with Build, Scale, and Convert agents or build custom agents.

Nvidia Releases New Open Transcription Model

Nvidia has released Nemotron 3.5 Transcribe, a free, open-weight speech-to-text model that can transcribe audio in up to 40 languages and translate 31 languages to and from English. It supports recordings up to one hour long and can distinguish up to eight speakers. The company also released version 2 of its Granary open-source speech dataset. The model is targeted at developers building meeting transcription and medical transcription applications. It runs on a single chip, from edge devices like Jetson AGX Thor to datacenter Blackwell chips. Nvidia claims it beats rivals on benchmarks like Hugging Face's Open ASR Leaderboard. However, it only works with recorded audio, not live streams, leaving real-time use to competitors like ElevenLabs. The release comes amid a wave of voice AI model launches from Mistral, ElevenLabs, and Salesforce.

• Nvidia released Nemotron 3.5 Transcribe, an open-weight speech-to-text model supporting 40 languages.
• The model can translate 31 languages to and from English and handle recordings up to one hour long.
• Nvidia also released version 2 of its Granary open-source speech dataset.

Block Unveils Human-Sounding Voice Model Dessa

Block, the fintech company led by Jack Dorsey, has released a new voice model called Dessa from its research lab. The model is trained fully on human speech and can run locally on a consumer-grade MacBook M4 Max. Dessa is a full-duplex model, enabling simultaneous listening and speaking, similar to OpenAI's GPT-Live. Unlike many voice models that rely on text pipelines and voice tags, Dessa embeds emotion directly into the model, as demonstrated in samples where the bot conveys tiredness or emotional distress. The company claims this leads to more natural interactions, addressing the loss of emotional information in text-based approaches. Currently, Dessa is not publicly accessible, and Block has invited safety researchers to test the model for potential risks, such as deceptive voice agents. This development could enable more cost-effective voice experiences for businesses, building on Square's existing voice AI for ordering.

• Block released a new voice model called Dessa.
• The model runs locally on a MacBook M4 Max.
• Dessa is a full-duplex model that can talk and listen simultaneously.

AI is changing customer service jobs, for better or worse

Artificial intelligence is transforming customer service roles, with AI tools handling routine tasks and administrative work, allowing human agents to focus on complex cases. However, employees report increased workloads and new challenges as AI errors create extra work. Companies like Virgin Media O2 and ArvatoConnect are adopting AI agents, with ArvatoConnect aiming for 70% of interactions handled by technology. Experts warn of work intensification and potential burnout as agents handle tougher cases, while metrics shift toward customer outcomes rather than speed. The article highlights the mixed impact on employees, with some seeing benefits and others feeling overworked.

• AI tools are taking over routine customer-service tasks, allowing humans to handle complex cases.
• Virgin Media O2 launched an AI voice agent in September 2026.
• ArvatoConnect aims for 70% of interactions to be handled by technology.

Voice AI Leaders Say Tech Hasn't Had Its ChatGPT Moment Yet

At the HumanX conference, executives from PolyAI and Otter discussed the current state and future of voice AI. PolyAI's CTO Shawn Wen believes that while full-duplex models have been achieved, voice AI still lacks a 'ChatGPT moment' due to the need for faster reasoning and more natural conversations. Otter's CMO Alex Gay emphasized the importance of speaker identification and accurate transcription as foundational for automation, and highlighted work on digital twins with emotive voices. Both executives stressed the need for transparency, ensuring users know when they are interacting with AI. They identified challenges such as ASR accuracy and language understanding, which are critical for building trust and effective downstream actions.

• PolyAI CTO Shawn Wen stated voice AI hasn't reached its 'ChatGPT moment' despite full-duplex models.
• Otter CMO Alex Gay emphasized speaker identification and accurate transcription as critical for automation.
• Otter is developing digital twins that can represent people in meetings and require emotive voice output.

Dialog-RSN-1 is now on AWS SageMaker

PolyAI announces that Dialog-RSN-1 is now available on AWS SageMaker, enabling deployment of their voice AI model on the AWS platform.