ElevenLabs launches Conversational AI 2.0 voice-assistants that know when to pause, talk, and take turns

AI advances are accelerating for businesses. This is especially true for speech and voice AI models.

A case in point: Today ElevenLabs is the well-funded voice AI sound effects startup founded and funded by former Palantir Engineers. Conversational AI 2.0is a significant upgrade of its platform to build advanced voice agents, which can be used for enterprise use cases such as customer service, call centers, outbound sales, and marketing.

The update introduces new features that are designed to create more intelligent, natural, and secure interactions. It is well-suited for enterprise applications.

ElevenLabs’ commitment towards rapid development is reflected in the launch of the new platform just four months after its debut. It also comes a day after Hume, a rival voice AI startup, launched EVI 3, based on a turn-based AI voice model.

The launch comes just four months after the debut of the original platform, reflecting ElevenLabs’ commitment to rapid development. It also comes a day after Hume launched its own new voice AI model based on turns called EVI 3. These declarations were, of course, premature.

Jozef Marko, from ElevenLabs’ engineering team, says that Conversational AI 2.0 has a much higher quality than its predecessor. It sets a new standard in voice-driven experiences.

Enhancing naturalistic speech.

The state-of-the art turn-taking model is a key feature of Conversational AI 2.0. This technology is designed for human conversation and eliminates awkward pauses or interruptions. By analyzing real-time conversational cues such as hesitations and filler phrases, the agent will be able to understand when to speak or listen. This feature is especially relevant for applications like customer service, in which agents must balance quick answers with the natural rhythms and flow of a conversation.

Multilingual support

Conversational AI 2.0 introduces integrated language detection to enable seamless multilingual conversations without the need for manual setup. This feature allows the agent to recognize the language of the user and respond appropriately within the same conversation.

This feature is designed for global enterprises that want to provide consistent service to diverse customer bases. It removes language barriers and fosters more inclusive experiences.

Enterprise-grade

The Retrieval-Augmented Generation system (RAG) is one of the most powerful additions. This feature allows AI to access external databases and retrieve relevant data instantly, while maintaining minimal latencies and strong privacy protections.

In healthcare settings, for example, a medical assistant agent could retrieve treatment guidelines directly from a database of an institution without delay. Agents can access the latest product information from internal documentation in order to provide better customer service.

Multimodality and alternate Personas

ElevenLabs’ new platform supports multimodality. Agents can communicate using voice, text or a combination. This flexibility reduces the burden on developers as agents can be defined only once to operate across multiple communication channels.

Conversational AI 2.0, which enhances agent expressiveness by allowing agents to switch between multiple personas, allows for multi-character mode. This capability can be useful in scenarios like creative content development, simulations for training, or customer engagement campaigns.

Using Conversational AI agents to initiate multiple outbound phone calls simultaneously is a great way for organizations to conduct surveys, alerts and personalized messages.

This new feature is designed to increase both reach as well as operational efficiency. It offers a more scalable option to manual outbound efforts.

Enterprise-grade pricing plans and standards

Beyond features that enhance communication, Conversational AI 2.0 puts a strong focus on compliance and trust. The platform is HIPAA compliant, which is a requirement for healthcare applications requiring strict privacy and data security. It also supports optional EU-based data residency, aligning itself with data sovereignty requirements across Europe.

ElevenLabs enhances these compliance-focused capabilities with enterprise-grade reliability and security. Conversational AI 2.0, designed for high availability and integration of third-party systems is a reliable and secure choice for businesses operating within sensitive or regulated environments. Here are the subscription plans that include Conversational AI. ElevenLabs website currently lists:

• : $0/month. Includes 15 minutes, four concurrency limits, requires attribution, and does not require commercial licensing.
• Starter: $5/month, includes 50 minutes, 6 concurrency limit.
• Create: $11/month, (discounted to $22), includes 500 minutes, 6 concurrency limits, and $0.12 for each additional minute.
• Pro: $99/month, includes 1,100 minutes, 10 concurrency limit, ~$0.11 per additional minute.
• Scale: $330/month, includes 3,600 minutes, 20 concurrency limit, ~$0.10 per additional minute.
• Enterprise: $1,320/month includes 13,750 minutes and 30 concurrency limits, $0.096 for each additional minute.

As stated in the article, a new chapter in realistic and naturalistic AI voice interaction

has been added. The company’s video introducing its new releasesays, “The potential for conversational AI is greater than ever.” The time to build has come.

ElevenLabs’ Conversational AI 2.0 aims to provide enterprises with the tools and infrastructure to create intelligent, context-aware, voice agents that will elevate the standard of digital interaction.

ElevenLabs encourages those who are interested to explore the documentation, visit the Developer Portal, or contact the sales team to learn more about how Conversational AI 2.0 will enhance their customer experience.