Google, Microsoft, and Salesforce have united to establish Agentic Resource Discovery, or ARD. This new open standard enables AI agents to autonomously locate and connect with tools across corporate software ecosystems. Announced in mid-June, the specification aims to define the future of workplace automation independent of chatbot-centric models.
ARD utilizes ai-catalog.json manifests to create a federated registry system. AI agents can now perform secure, intent-based searches to discover necessary resources without custom integrations. Organizations maintain control by publishing capabilities under their own domains, ensuring no single entity owns the network. The protocol operates under Apache 2.0 licensing with coordination from the Linux Foundation.
The supporting coalition includes major infrastructure players like Cisco, NVIDIA, Databricks, and Snowflake. Notably absent are OpenAI and Anthropic. While those firms focus on conversational interfaces, ARD prioritizes autonomous agent integration within established platforms like Google Workspace and Microsoft 365. The standard is designed to complement existing protocols, including Anthropic’s Model Context Protocol, rather than replace them.
For investors, ARD represents a strategic defense of subscription revenue models. By making their ecosystems the default destination for autonomous agents, these tech giants ensure continued relevance in an automated economy. Tools lacking ARD-compatible manifests risk becoming invisible to future AI workflows. This is a pure software standards play with no blockchain components, focusing entirely on reducing enterprise fragmentation and securing long-term platform utility.