Google, Microsoft, and Salesforce have jointly announced a new open standard called Agentic Resource Discovery (ARD).
Simply put, ARD is like a universal 'Yellow Pages' for AI. As companies increasingly use numerous specialized AI agents and digital tools, they often get scattered across different systems. ARD creates a standardized catalog, making it easy for AI agents to find, verify, and connect with the right tool or data source, no matter where it is. This is a direct response to the 'super-agent' approach favored by companies like OpenAI, which aims to create a central AI platform that controls all other apps, potentially leading to a 'lock-in' effect.
The alliance behind ARD is instead betting on an open ecosystem. Their strategy is to make their own enterprise applications—like Microsoft Copilot, Google Gemini, and Salesforce CRM—the primary gateways for using AI. By controlling the 'discovery layer' through an open standard, they guide AI usage toward their software ecosystem, rather than letting a single AI model dominate.
So, why is this happening now? Three key factors created the perfect conditions for ARD's emergence.
First, the issue of cost. GitHub recently switched its Copilot service to usage-based billing. This means that when an AI agent spends time searching for and calling various tools, the cost adds up. ARD helps reduce this by making the discovery process more efficient, directly addressing companies' concerns about rising AI expenses.
Second, the need for security and trust. A recently discovered security flaw in the Model Context Protocol (MCP), a related technology, highlighted the risks of unverified connections between agents and tools. ARD incorporates a verification process, creating a trusted pathway for these interactions and building a more secure foundation for the agent ecosystem.
Third, the sheer growth in demand. Market analyst Gartner predicts that by the end of 2026, 40% of enterprise applications will include specialized AI agents. Without a common standard for discovery and governance, this rapid growth would lead to chaos and bottlenecks. ARD provides the necessary plumbing to manage this complexity, ensuring that the growing world of AI agents can work together smoothly and securely.
- AI Agent: A smart software program that can perform tasks on its own to achieve a specific goal. For example, an agent could book a trip by finding flights, reserving a hotel, and adding it to your calendar.
- API (Application Programming Interface): A set of rules that allows different software applications to communicate with each other. It's like a menu in a restaurant that lets you order food from the kitchen.
- Lock-in Effect: A situation where a customer becomes dependent on a single company's products or services and cannot easily switch to a competitor without substantial costs or inconvenience.