KEY POINTS
Accenture introduces Trusted Agent Huddle, a new capability to manage AI agents within large organizations, enhancing collaboration and control.
The platform aims to transform cross-functional workflows by enabling interoperability among major enterprise software systems.
It uses open standardization protocols and a proprietary algorithm to ensure agent performance and trust.
"With the proliferation of AI agents across the enterprise, trust is the only limit to AI reaching its full potential."
While traditional code offered predictable, deterministic outcomes, AI thrives on nuance, connecting disparate systems with unprecedented power. But that same variability understandably makes enterprises uneasy. Seeking to bridge this gap between AI's potential and the need for reliable control, Accenture has introduced Trusted Agent Huddle, a new capability designed to manage the burgeoning ecosystem of AI agents within large organizations.
Orchestration: Trusted Agent Huddle operates within Accenture’s existing AI Refinery platform. The company says its primary function is to enable secure and seamless collaboration between AI agents developed by different enterprise users and across various technology platforms, allowing companies to select and manage agents for specific tasks.
Bridging the vendor gap: Accenture highlights the platform's ability to foster interoperability across a wide array of major enterprise software partners. The goal is to move beyond optimizing isolated processes within single systems toward transforming entire, cross-functional workflows.
Built in predictability: Addressing the inherent trust challenge with complex AI systems, Accenture says Trusted Agent Huddle utilizes open standardization protocols like Agent2Agent and Model Context Protocol for onboarding agents. The system employs a proprietary algorithm to evaluate and align agent performance, described as a certification process laying groundwork for a future "agent trust score."
Lan Guan, Chief AI Officer of Accenture, said in the announcement, "With the proliferation of AI agents across the enterprise, trust is the only limit to AI reaching its full potential."
"Collaboration and integration across platforms and systems is key to running AI agents that can reason and complete complex tasks."
Powering the platform: Accenture said its AI Refinery was built on NVIDIA AI Enterprise and the underlying technology involves a significant collaboration with the chip maker. The new Trusted Agent Huddle can leverage the NVIDIA Agent Intelligence toolkit, which facilitates connectivity between different agents, tools, and data sources. "Collaboration and integration across platforms and systems is key to running AI agents that can reason and complete complex tasks," said Justin Boitano, Vice President of Enterprise AI Software Products at NVIDIA, in the release.
Real-world test: Logistics giant FedEx is already working with Accenture and NVIDIA to explore uses for Trusted Agent Huddle. "The Trusted Agent Huddle enables agents from different platforms to collaborate as one team and will help FedEx seamlessly engage and orchestrate information across the supply chain to accelerate innovation," said Sriram Krishnasamy, Chief Transformation Officer and Chief Digital & Information Officer at FedEx Corporation. Karthik Narain, Group Chief Executive—Technology and chief technology officer at Accenture, added that the platform helps enterprises like FedEx "future proof their innovation and AI investments."
Broader AI push: This platform launch follows other recent AI-focused moves by Accenture, including an investment through Accenture Ventures into GoUpscale, an AI-powered content engagement technology company focused on the wealth management sector—signaling a multi-faceted strategy for Accenture that combines platform development with targeted investments.