NIST Selected Two Use Cases for Agentic AI Project

The federal agency has identified focus areas for its upcoming work following a period of public and stakeholder feedback.

Updated on Sept. 30, 2026 in Artificial Intelligence

Isometric editorial illustration featuring two interlocking geometric polymer forms and a clean silicon wafer, representing research focus areas. — The National Institute of Standards and Technology has narrowed its agentic AI research to two specific focus areas to help standardize evaluations for autonomous software systems. AI Illustration. Upload story photo >

Live Poll

Should federal agencies prioritize stakeholder feedback when developing new artificial intelligence standards?

The National Institute of Standards and Technology (NIST) has officially selected two use cases for its upcoming research project on agentic artificial intelligence. This decision follows the agency's issuance of a concept paper in February 2026 and a subsequent review of feedback from the agentic AI ecosystem.

Why it matters

By narrowing its scope to these two specific use cases, the agency aims to build consensus among industry stakeholders and research groups. This step is a critical precursor to establishing standardized evaluations for AI agents capable of autonomous decision-making.

NIST has defined the scope for its agentic AI project as two specific use cases. The agency reached this determination by analyzing observations and feedback collected from the agentic AI community following the February 2026 concept paper.

The players

National Institute of Standards and Technology

A federal agency that develops technical standards and benchmarks for emerging technologies, including cybersecurity and artificial intelligence.

The details

Agentic AI refers to software systems designed to perceive their environment, reason, and execute multi-step actions to achieve objectives without constant human guidance. The agency is now moving to align these technical developments with government standards. This process relies on integrating insights from public comments to ensure the project addresses practical implementation challenges within the AI ecosystem.

February 2026: The agency issued a concept paper regarding the project.
September 30, 2026: NIST finalized the selection of the two use cases.

The Tech Race

This project follows the pattern established by the NIST AI Risk Management Framework by seeking to formalize evaluation criteria for a new class of AI capability. The agency is now positioning itself to lead the development of consensus-based standards for increasingly autonomous software agents.

The agency's selection of these two use cases serves as a signal to industry developers regarding where to focus their internal compliance and safety efforts. As these standards mature, they will likely influence the architectural requirements for any AI agents deployed within regulated professional sectors.

The takeaway

The agency's shift from broad concept paper to defined use cases suggests a transition toward concrete regulatory frameworks for autonomous agents. Observers should track the next formal release from NIST to identify the specific industry sectors prioritized for these initial benchmarks.