The new IVA Studio platform integrates autonomous agentic workflows with a modular skills framework to standardize enterprise procurement operations.
Ivalua released IVA Studio on June 11, 2026, introducing a centralized artificial intelligence architecture designed to automate complex procurement and supply chain operations. The platform centers on a primary agent, IVA, which functions as an interface across the company’s existing source-to-pay ecosystem.
The system operates by accessing the underlying Ivalua platform as both a knowledge repository and a functional toolset. This integration allows the agent to execute tasks such as contract retrieval, supplier risk assessment, and invoice reconciliation without requiring separate configuration for each module. By leveraging the platform’s native data, the agent maintains operational continuity across sourcing and accounts payable workflows.
David Khuat-Duy, founder and chief AI officer at Ivalua, stated that the architecture avoids the latency associated with traditional agent-building processes.
Procurement teams have typically had to spend time building and configuring AI agents before seeing any results. Ivalua’s approach is entirely different: IVA accesses the Ivalua platform as its toolset and source of knowledge, so procurement can start getting value from day one, within a framework of governance that’s enforced by design.
The technical implementation relies on a skills framework that mirrors structures found in advanced machine learning research. Organizations can encode specific negotiation tactics, compliance protocols, and sourcing methodologies as reusable skills, which the agent then applies to various tasks. This modularity allows for the codification of institutional knowledge into a shared operational model.
Memory management within the system enables the agent to refine its performance based on historical interactions. The software supports the Model Context Protocol, providing the flexibility to connect with external systems and diverse large language models. This model-agnostic design permits IT departments to select or swap underlying models based on specific performance requirements or data privacy constraints.
The integration of the Model Context Protocol allows the agent to maintain a consistent state across heterogeneous data sources. By standardizing how the agent interacts with external APIs, Ivalua ensures that the system can pull real-time market data or supplier performance metrics without custom middleware. This capability is critical for maintaining the accuracy of the agent’s decision-making process in volatile supply chain environments, where data latency often degrades model performance.
Governance remains a primary design constraint, with the system inheriting the specific access permissions of the human user currently logged into the platform. Every autonomous action performed by the agent is recorded in an immutable audit trail, ensuring that all background processes remain within established internal controls. This structure addresses critical requirements for transparency in financial and supply chain operations where regulatory compliance is mandatory.
The shift toward agentic workflows reflects a broader transition in enterprise software, where the focus moves from static automation to dynamic, context-aware decision support. By enabling the agent to coordinate temporary sub-agents for complex tasks, Ivalua aims to reduce the cognitive load on procurement professionals. This approach allows teams to delegate repetitive analytical processes while retaining human oversight for strategic relationship management and high-level decision-making.
Franck Lheureux, chief executive officer of Ivalua, noted that the objective is to enhance human productivity rather than replace it. The system’s ability to operate autonomously within a strictly defined permission boundary provides a mechanism for scaling procurement capacity without compromising internal risk management. As the platform matures, the integration of these agentic capabilities will likely serve as a benchmark for how enterprise resource planning systems incorporate autonomous, model-agnostic intelligence.
Future development cycles will likely focus on expanding the library of reusable skills and improving the agent’s ability to handle multi-step, cross-functional procurement events. The company plans to monitor the efficacy of these autonomous workflows across its network of over 500 brands to further refine the underlying skill-based architecture and improve overall system reliability.
