ElevenLabs has launched a new enterprise-focused feature called Agent Transfer, aimed at enabling artificial intelligence agents to communicate and exchange conversation data seamlessly.
Highlights
Announced via a post on X , the update is part of the company’s broader Conversational AI framework and reflects an effort to streamline AI-driven workflows across departments.
The core functionality of Agent Transfer allows conversations to be passed between AI agents when specific conditions are met—such as escalation scenarios or limitations in an agent’s knowledge base.
Rather than escalating directly to a human representative, the system enables a smooth handover to another specialized AI agent. Importantly, the full conversation history is retained, allowing the receiving agent to maintain continuity and context in the interaction.
The feature is currently available to enterprise clients. However, ElevenLabs has not specified whether it will be offered as part of existing plans or as a standalone service. Implementation guidance is available through developer documentation on the company’s support page.
Designed to address fragmentation in enterprise AI systems, Agent Transfer aims to eliminate the data silos that often arise when multiple AI agents are deployed in isolation.
By supporting inter-agent communication, the system promotes interoperability and allows information to flow fluidly across functions.
This design contrasts with centralized data architectures adopted by some AI companies, instead favoring modular, agent-based collaboration to scale AI infrastructure.
In practical terms, the feature can significantly enhance customer support workflows. For instance, a company may operate several AI agents, each specialized in areas like general inquiries, technical troubleshooting, or billing.
With Agent Transfer, a customer query that exceeds the scope of a general inquiry agent can be seamlessly redirected to a more appropriate agent, without the user needing to repeat information or restart the conversation.
Structured Handoff for Smoother Interactions
The Agent Transfer mechanism is intended to improve the accuracy and efficiency of AI-supported interactions.
By allowing agents to identify when an issue falls outside their scope and redirect the conversation accordingly, the system helps ensure users receive timely and relevant support. For businesses, this also optimizes internal resource allocation and reduces redundant task routing.
GibberLink: AI-to-AI Communication
In a related initiative, ElevenLabs also introduced GibberLink during its London Hackathon. Developed by Boris Starkov and Anton Pidkuiko, GibberLink is a protocol that enables AI agents to communicate using structured sound waves instead of human language.
Built on the open-source ggwave library, this system allows agents to recognize when they are interacting with another AI and switch to a more efficient non-verbal mode of communication.
According to the developers, GibberLink can reduce computational overhead by up to 90% and increase communication speed by approximately 80%.
While still in an experimental phase, this approach highlights the potential for optimizing inter-agent communication beyond traditional text or voice interfaces.
Enterprise Implications and Evolving AI Architectures
The introduction of Agent Transfer and experimental tools like GibberLink reflects a growing industry shift toward more autonomous, adaptive AI systems.
For enterprises, these capabilities could lead to AI ecosystems that not only handle discrete tasks but also collaborate dynamically, adapting to changing user needs and complex service environments.
This agent-based architecture supports distributed task handling based on expertise, potentially reducing bottlenecks and reliance on centralized systems or human intervention.
ElevenLabs positions this model as a way to improve responsiveness and personalization in AI-driven operations.
Transparency Considerations in AI Communication
While these advancements offer practical benefits, they also raise questions around transparency and oversight.
GibberLink’s use of non-human communication modes, for example, highlights a challenge in ensuring AI interactions remain interpretable to human supervisors.
As AI agents gain more autonomy, maintaining visibility into their operations and ensuring accountability will be critical for responsible deployment.
Enterprise Adoption
The effectiveness of Agent Transfer will depend largely on how organizations structure their AI workflows and train agents for specialization.
ElevenLabs has yet to release client feedback or performance metrics, but the rollout reflects the company’s continued investment in building interoperable AI solutions for enterprise environments.
