Samaya AI announced it has introduced a new architecture, the Agent Control Plane (ACP), positioning it as a control and governance layer for building and operating finance-focused AI agents that reason in real time within a specialized investment context. The Mountain View-based company also disclosed new investment from NVentures, NVIDIA’s venture capital arm, and Databricks Ventures, which it plans to use to extend development of ACP and expand customizable agents for financial workflows.
The company framed ACP as an answer to the limitations it says generic AI agents face in financial services, including context constraints, error compounding, and performance degradation as the number of tools and data sources increases. Samaya said its approach is built to help users design, run, and govern agents tailored to specific financial use cases through natural-language instructions and iterative refinement, while preserving transparency into how outputs and conclusions are generated.
Samaya said ACP incorporates multiple components it has developed for financial reasoning, including an integrated planning, execution, and memory stack, a reasoning module, and modules for context management and tool optimization. The company argued that the combination is intended to deliver accurate, auditable results for high-stakes workflows that require synthesizing information from large datasets, enterprise systems, and external content sources.
Samaya said it is already operating in production with more than 10,000 professionals at one of the world’s largest banks, and that additional financial services clients are deploying ACP at enterprise scale. The company said use cases include earnings analysis as well as broader scenario analysis, modeling, and economy-wide research tasks, with agents designed to blend internal enterprise data with external sources to produce institution-specific outputs aligned with user context.
Maithra Raghu, CEO and founder of Samaya AI, said finance agents must convert global information into real-time decisions and remain reliable, auditable, and precise while reasoning across large volumes of data and tooling. She said Samaya’s ACP is intended as a blueprint for building and operating next-generation finance agents and that the company is already seeing results from investment professionals using the architecture.
In connection with the investment, Samaya said it aims to deepen ACP’s capabilities and expand its positioning as an “institutional intelligence layer” for financial services, focusing on agent customization and governance for investment workflows.
KEY QUOTE:
“AI Agents for finance have to transform global information into real-time decisions. They must be reliable, auditable, and extraordinarily precise while reasoning across millions of datapoints and hundreds of tools. Generic AI Agents break down under this complexity — hitting context limits, compounding errors, and degrading as tools scale. Samaya’s ACP is a blueprint for how next-generation AI Agents should be built and run, and we’re already seeing meaningful results from investment professionals using ACP across the globe.”
Maithra Raghu, Chief Executive Officer and Founder, Samaya AI
