Genesys has announced the launch of the Genesys Cloud™ Agentic Virtual Agent, marking the industry’s first enterprise-grade solution powered by Large Action Models (LAMs). The technology is designed to move beyond simple conversation, enabling AI to autonomously execute end-to-end resolutions across back-office systems and multi-step workflows.
The announcement comes at a critical juncture for the customer experience (CX) industry. While 46% of business leaders identify increasing self-service success as a top-three priority for 2026, current success rates for self-service interactions hover at just 22%, according to Gartner research.
Closing the action gap
While previous generations of virtual agents relied on Large Language Models (LLMs) to improve conversational tone, they often struggled with execution. LLM-based bots frequently fail when tasks require navigating multiple systems—such as billing, CRM, and service operations—or when customer needs shift mid-interaction.
The Genesys Cloud Agentic Virtual Agent addresses this by utilising the Scaled Cognition APT-1 LAM. Unlike text-focused models, this action-grounded AI is built for deterministic execution, allowing it to plan and carry out tasks across enterprise systems while adhering to strict company policies.
“Autonomy in customer experience only works when it’s built on trust, transparency, and control,” said Olivier Jouve, Chief Product Officer at Genesys. “We’re enabling AI to reason, plan, and safely take action across systems… giving organisations a responsible way to move beyond conversations.”
Governance and global adoption
To mitigate the risks of hallucinations or policy drift, Genesys has embedded guardrails into the platform. Through the Genesys Cloud™ AI Studio, enterprises can define specific permissions and behaviors, ensuring every AI-driven action is auditable and explainable.
Several global organizations have already begun exploring the technology, including:
- M&T Bank and Banco Pichincha
- DNB, Norway’s largest financial services group
- A Fortune 50 North American retailer and a global Fortune 500 healthcare company
“In the enterprise, 80% accurate is 100% useless for automation,” noted Dan Roth, co-founder and CEO of Scaled Cognition. “The foundation of trustworthy automation is super-reliability. By combining our Large Action Models with the governance of Genesys Cloud, we deliver a more reliable AI stack.”
Future interoperability
Looking forward, Genesys plans to expand the capabilities of the Agentic Virtual Agent through support for open standards such as Agent-to-Agent (A2A) and Model Context Protocol (MCP). This will allow Genesys agents to collaborate securely with other AI agents and third-party enterprise systems, maintaining context across complex, multi-step customer journeys.
As the market shifts toward Agentic CX, industry analysts suggest this move is necessary for the next phase of digital transformation. Hayley Sutherland, Research Manager at IDC, noted that the focus on combining autonomous execution with rigorous governance “reflects the direction the market needs to move to make self-service both more effective and trustworthy.”
