# Agentic AI Foundation: A Community Unifying AI Tools and Standards ## What Is the Agentic AI Foundation? The **Agentic AI Foundation (AAIF)** is an open community, founded in 2025 by the [Linux Foundation](https://www.linuxfoundation.org/), to provide a vendor-neutral approach for frameworks, standards, and practices as Agentic AI systems move from experimentation into real-world production. AAIF develops open standards and frameworks through **6 Working Groups** to collaborate on focused initiatives, share best practices and expertise, and drive practical outcomes across the AI ecosystem for building production-ready agentic AI systems. Its key members currently include **OpenAI, Google, Microsoft, AWS, Anthropic, Block, Bloomberg, Cloudflare**, and many more. ## The 6 Working Groups Each working group owns standards and best practices for one part of the agentic stack: | Working Group | What it defines | | :--- | :--- | | **Taxonomy & Landscape** | A shared glossary of agentic AI terms and an ecosystem map so every other group works from common definitions. | | **Accuracy & Reliability** | What reliability, accuracy, and consistency mean for autonomous systems — including failure management and SLAs. | | **Agentic Commerce** | Agent discovery, negotiation, payment authorization, and protocols for trustworthy autonomous transactions. | | **Governance, Risk & Regulatory Alignment** | Shared frameworks aligning agentic innovation with legal, ethical, and regulatory expectations (e.g., the EU AI Act). | | **Identity & Trust** | Portable identity and dynamic trust for agents — delegation protocols, cross-domain identity, and permission flows. | | **Security & Privacy** | Industry benchmarks for secure agentic operations, security-by-design, and adversarial testing methodologies. | ## Key Questions ### Q1) What does the Agentic AI Foundation actually produce? Open standards, best practices, and reference frameworks — organized across 6 working groups (Taxonomy & Landscape, Accuracy & Reliability, Agentic Commerce, Governance/Risk/Regulatory Alignment, Identity & Trust, and Security & Privacy) plus 5 hosted projects. ### Q2) What are the key standards available as part of AAIF? - [MCP](https://aaif.io/projects/model-context-protocol): a protocol for data integration between LLM applications and external data sources, donated by [Anthropic](https://www.anthropic.com/news/donating-the-model-context-protocol-and-establishing-of-the-agentic-ai-foundation). - [AGENTS.md](https://aaif.io/projects/agents-md): an open standard for guiding coding agents, donated by [OpenAI](https://openai.com/index/agentic-ai-foundation/). - [Agent2Agent (A2A)](https://aaif.io/projects/agent2agent): an open protocol designed to enable interoperability between AI agents, donated by [Google](https://aaif.io/blog/a2a-joins-aaif). ### Q3) Why this matters for architects? For teams building production agentic systems today, AAIF serves not only as a shared community but also as a unified place for common standards, best practices, and tools, including: - [AAIF Agentic AI Projects List](https://aaif.io/projects) - [AAIF Agentic AI Landscape](https://landscape.aaif.io/) - [AAIF Blogs](https://aaif.io/blog) covering key topics such as: - [Where A2A fits in the open agent ecosystem](https://aaif.io/blog/where-a2a-fits-in-the-open-agent-ecosystem) - [Enterprise MCP gateway security in practice with agentgateway](https://aaif.io/blog/enterprise-mcp-gateway-security-in-practice-with-agentgateway) - [Writing an effective AGENTS.md](https://aaif.io/blog/writing-an-effective-agents-md) ### Q4) Is there a visual interface for exploring Agentic AI standards, tools, and technologies? Yes. Similar to the [CNCF Cloud-native Landscape](https://landscape.cncf.io/), AAIF has published an [Agentic AI Landscape](https://landscape.aaif.io/), shown below: {/* Landscape embed view */}