Consumer loyalty is tested in the moments that matter most. A customer spots an unauthorized charge and wants it resolved on the spot; a policyholder calls after a fire desperate for guidance while their home still smolders; or a patient waits for treatment approval with each day of delay carrying real consequences. These are not times for apologies or vague promises. People expect progress and empathy delivered as swift, dependable action from both artificial intelligence (AI) and human agents. 

Genesys is focused on meeting this standard with agentic orchestration. Today we announced a new partnership with Scaled Cognition that will deepen our capabilities to build AI that turns intent into outcomes.  

Together, we’re charting the future where AI agents don’t stop at conversations. They will take the next action, resolve the issue, and deliver results reliably and under enterprise-grade governance. By integrating the Scaled Cognition APT-1 large action model into the Genesys Cloud™ platform, organizations will be able to deploy autonomous virtual agents that can complete workflows in real time, adhere to compliance policies and deliver customer experiences where resolutions define success. 

This is the next evolution of agentic orchestration — AI systems built to carry intent through to completion by turning complex workflows into predictable, repeatable outcomes. 

Turning Words Into Actions with LAMs 

Traditional large language models (LLMs) are trained from text and designed to converse about next steps rather than execute them. They can explain how a claim will be processed or when an account might be updated, but they can’t reliably perform the outcome.   

Large action models (LAMs) extend the conversational strengths of LLMs by adding the ability to decide, act and verify. This will enable richer, more complex experiences and empower virtual agents to carry a workflow to completion across multiple systems and policies.  

LAMs are trained to take actions by combining conversation with the policies and rules that govern real-world workflows. Instead of just predicting the next word, they also decide on the correct step, call the right API, enforce the right policy or escalate when needed.   

Because execution is part of their training, they can behave more deterministically, reduce variance and deliver more reliable outcomes. The impact will be stronger containment, faster resolution and the ability to eliminate hallucinations. Businesses gain efficiency and compliance assurance while customers see faster and better results. 

Consider a billing-resolution scenario: A customer disputes a charge and reaches out for help. An LLM-based agent responds with empathy, guides the customer through the process and outlines what will happen next.  The actual refund, however, relies on downstream processes and follow-up steps.  

A LAM-based agent would expand these capabilities by authenticating the account, posting the credit, verifying the transaction and confirming the resolution within the same interaction. In this case, a solution could be delivered immediately. It closes the loop for the customer and executes the downstream operational steps all within the same interaction. 

The Genesys and Scaled Cognition Advantage 

Anyone can license a traditional LLM. Beyond access to models, what makes Genesys different is how we orchestrate them inside our unified platform that’s purpose-built for customer experience (CX). 

Genesys Cloud already supports a broad range of options, including Genesys-built, open-source and frontier models through Amazon Bedrock. This allows enterprises to innovate with models that best align with their goals while maintaining safety, control and performance.  

Scaled Cognition LAMs build on that foundation. They combine the conversational strength of LLMs with action-driven intelligence. By pairing the orchestration power of Genesys Cloud with the Scaled Cognition purpose-built models for customer experience, enterprises will have access to autonomous agents that can operate with even more precision and control.  

Organizations will be able to design virtual agents that can reliably execute complex tasks, coordinate seamlessly with human teams and uphold compliance standards. What makes this distinctive is how intricate business logic is translated into deterministic behaviors to produce outcomes that are consistent, auditable and aligned with enterprise policies, including: 

  • Governance as Architecture, Not an Afterthought: Genesys Cloud is built with embedded governance, and guardrails within Genesys Cloud AI Guides help keep actions aligned with policy, minimize hallucinations and lessen drift. With deterministic LAMs from Scaled Cognition those protections will grow stronger. Simulation-based certification, runtime policy enforcement and transparent action logs will be able to give enterprises assurance before, during and after deployment.
     
  • Reliability that Scales: Traditional LLMs often require human spot-checking. Deterministic models reduce the variance and will allow Genesys to add new automated compliance layers, including audits, to help prove consistency; monitoring that flags exceptions; and certification that validates workflows before they go live.
     
  • Orchestration from the Center: Inside Genesys Cloud, we expect that the Scaled Cognition LAM will work through Genesys CloudTM AI Studio and AI Guides. They won’t float as add-ons. They will drive multistep workflows within the same platform that supports voice, chat and digital channels.
     
  • Context that Connects Journeys: Genesys Cloud links every action to its unified data fabric. Each decision sits in the stream of customer history, events and channels. Others might stitch together point solutions, but Genesys delivers journeys that are connected from start to finish. 

Genesys is powering the strategic platform for agentic orchestration to help enterprises create autonomous CX agents that can act decisively and govern transparently across systems.  

Working together with our extensive ecosystem of partners, which now includes Scaled Cognition, we give organizations the ability to rapidly innovate and keep compliance front and center. The addition of the Scaled Cognition APT-1 model strengthens this vision and adds deeper intelligence to the Genesys Cloud platform, so businesses can deploy agentic agents with confidence at scale.  

Three Moves to Prepare Your Business for Agentic Orchestration 

The path to LAM adoption benefits from deliberate preparation. Three actions organizations should consider are: 

  • Audit for Sticking Points: Pinpoint where customer journeys break down. Compliance checks, policy enforcement and multistep approvals are prime candidates for action-driven orchestration
  • Governance as a Trust Builder: Strong oversight builds confidence. Use simulation to validate behavior, enforce escalation rules and track runtime performance. Deterministic models make certification and auditing straightforward
  • Think in Ecosystems, Not Silos: Adopt standards that ensure flexibility. Genesys Cloud will be MCP-ready; combining it with the Scaled Cognition models prepares enterprises for ecosystems that integrate across platforms and vendors. 

Fully Autonomous Journeys Are on the Horizon 

In the next year, we expect to see fully autonomous customer journeys in production; and Genesys is mapping the path into the agentic era. By partnering with Scaled Cognition and combining LAMs with governance, orchestration and data context, we will enable systems that act with accuracy, keep interactions compliant and complete customer journeys in real time.  

We believe our work with Scaled Cognition accelerates a future where AI will guide conversations and take ownership of outcomes. At Genesys, we believe the companies that embrace agentic orchestration now will define customer trust for the next decade. That’s why we’re building it into the core of our platform. 

To learn more about our collaboration with Scaled Cognition read our press release. 

Forward-Looking Statements  
Statements in this blog that are not historical or current facts are forward-looking statements that involve risks and uncertainties. Unless required by law, Genesys undertakes no obligation to update or revise any forward-looking statements to reflect circumstances or events after the date of this blog.