Four Protocols Are Building the Agentic Web's Foundation

The Infrastructure Shift That Changes Everything

The agentic web is being built on four foundational protocols that will determine which businesses thrive in the AI-first era. The Model Context Protocol (MCP) reached 97 million monthly SDK downloads in just over a year, signaling unprecedented adoption velocity. This matters because companies that understand these standards early will control how AI agents access their data, services, and content—creating structural advantages that competitors cannot easily replicate.

The parallel to the early web is exact but accelerated. Where HTTP and HTML took years to achieve critical mass, these agentic protocols are achieving similar adoption in months. MCP functions as the universal adapter, allowing any AI system to connect to any data source through a single interface. The Agent2Agent protocol (A2A) enables different AI systems to discover and collaborate with each other. NLWeb transforms websites into natural language interfaces, while AGENTS.md provides standardized guidance for AI coding agents. Together, they form the infrastructure layer that will support the next generation of business interactions.

The Governance Architecture That Prevents Fragmentation

The most significant development isn't the protocols themselves but the governance structure supporting them. The Agentic AI Foundation (AAIF), announced by the Linux Foundation on December 9, 2025, brings together eight platinum members who are direct competitors in the AI market: AWS, Anthropic, Block, Bloomberg, Cloudflare, Google, Microsoft, and OpenAI. This collaboration on infrastructure while competing on products mirrors the W3C's role in unifying the early web.

Jim Zemlin, Linux Foundation Executive Director, captured the moment: "We are seeing AI enter a new phase, as conversational systems shift to autonomous agents that can work together." This governance structure prevents the fragmentation that could have crippled the agentic ecosystem. Without AAIF, each major player would have developed proprietary standards, creating the M x N problem where businesses would need separate integrations for every AI platform and every tool. The economic inefficiency would have been staggering.

MCP: The Universal Adapter That Changes Integration Economics

MCP's adoption timeline reveals why this matters for business strategy. Launched by Anthropic on November 25, 2024, it gained native support in Claude from day one. By March 2025, OpenAI CEO Sam Altman announced support across OpenAI's products, stating: "People love MCP and we are excited to add support across our products." Google followed in April with Gemini support, and Microsoft joined the MCP steering committee in May 2025.

The strategic implication is clear: any business that makes its data, tools, or services MCP-accessible immediately becomes available to every major AI platform. This eliminates the need for separate integrations for ChatGPT, Claude, Gemini, and Copilot. For e-commerce companies, this means product catalogs, inventory systems, and order tracking become accessible to AI shopping assistants across all platforms simultaneously. The reduction in integration complexity translates directly to reduced costs and faster time-to-market.

A2A: Creating the Agent-to-Agent Economy

While MCP connects agents to tools, A2A connects agents to each other. Launched by Google on April 9, 2025, with over 50 technology partners, A2A grew to over 150 supporting organizations by July 2025, including enterprise heavyweights like Salesforce, SAP, ServiceNow, PayPal, Atlassian, Microsoft, and AWS. The protocol was donated to the Linux Foundation in June 2025, ensuring its vendor-neutral development.

The core innovation is the Agent Card—a JSON metadata document that serves as a digital business card for agents. Each A2A-compatible agent publishes an Agent Card describing its identity, capabilities, skills, and authentication requirements. When one agent needs help with a task, it reads another agent's card to understand what that agent can do, then communicates through A2A to request collaboration.

This creates what Google frames as: "Build with ADK, equip with MCP, communicate with A2A." The practical business application is transformative. Consider a customer service scenario where a billing question requires a refund. A customer service agent identifies the issue, passes context to a billing agent via A2A, which calculates the refund amount and hands off to a payments agent to process it. The customer sees one seamless interaction while three agents from different vendors collaborate through a shared protocol.

NLWeb: The Most Immediate Business Impact

For most businesses, NLWeb represents the most accessible entry point into the agentic web. Introduced at Microsoft Build 2025 on May 19, 2025, and developed by R.V. Guha—creator of RSS, RDF, and Schema.org—NLWeb builds directly on existing structured data investments. Microsoft's framing is deliberate: "NLWeb can play a similar role to HTML in the emerging agentic web." The NLWeb README puts it even more directly: "NLWeb is to MCP/A2A what HTML is to HTTP."

Every NLWeb instance automatically becomes an MCP server, meaning any website running NLWeb immediately becomes accessible to the entire ecosystem of MCP-compatible AI assistants and agents. Early adopters include Eventbrite, Shopify, Tripadvisor, O'Reilly Media, Common Sense Media, and Hearst—all content-rich websites that already invest heavily in structured data.

The business implication is straightforward: instead of users navigating search filters, AI agents can query websites directly using natural language. For example, an AI agent could query Tripadvisor's NLWeb endpoint: "Find family-friendly restaurants in Barcelona with outdoor seating and good reviews." The response comes back as structured Schema.org JSON, ready for the agent to present or act upon. Companies that have invested in Schema.org markup are already positioned for NLWeb adoption.

AGENTS.md: The Hidden Productivity Multiplier

AGENTS.md, while less visible to non-technical executives, represents a significant productivity multiplier. Emerging from collaboration between OpenAI Codex, Google Jules, Cursor, Amp, and Factory in August 2025, it has been adopted by over 60,000 open-source projects and is supported by tools including GitHub Copilot, Claude Code, Cursor, Gemini CLI, and VS Code. With GitHub reporting that Copilot now generates 46% of code for its users, standardized guidance for AI coding agents becomes essential quality control.

The file itself is simple—plain Markdown, typically under 150 lines, covering build commands, architectural overview, coding conventions, and testing requirements. Agents read it before making any changes, getting the same tribal knowledge that senior engineers carry. For development teams using AI coding tools, AGENTS.md ensures consistent output, reduces bugs, and cuts onboarding time for AI tools on new projects.

The Structural Winners and Losers

The emergence of these protocols creates clear structural advantages for specific categories of businesses. Major cloud and AI providers (AWS, Google, Microsoft, OpenAI, Anthropic) win as standards setters through AAIF governance, with early protocol adoption integrated into their core products. Enterprise software companies (Salesforce, SAP, ServiceNow, Atlassian) win through early A2A adoption, enabling AI agent integration into business workflows and creating new value propositions.

E-commerce platforms (Shopify, Etsy, Target, Walmart) win through commerce protocol development, with Shopify and Google co-developing the Universal Commerce Protocol (UCP) launched in January 2026, and OpenAI and Stripe co-developing the Agentic Commerce Protocol (ACP) powering Instant Checkout in ChatGPT. Content publishers and websites win through NLWeb adoption, making content more accessible to AI agents and potentially increasing visibility and usage.

The losers are equally clear. Traditional SEO-focused companies lose as the shift from search optimization to agent optimization (AAIO) requires fundamental strategy changes. Proprietary AI solution vendors lose as open standards reduce lock-in advantages and increase competition. Legacy integration platforms lose as agent-to-agent protocols could bypass traditional middleware for AI collaboration. Manual coding workflows lose as AI coding agents reduce demand for traditional programming approaches.

The Commerce Layer Emerging Above Infrastructure

Beyond the four foundational protocols, commerce-specific standards are building the transaction layer. Shopify and Google's Universal Commerce Protocol (UCP) and OpenAI and Stripe's Agentic Commerce Protocol (ACP) represent the next layer of specialization. These protocols enable AI agents to complete transactions directly, transforming how commerce happens.

The strategic implication is that businesses need to think in layers: infrastructure protocols (MCP, A2A, NLWeb, AGENTS.md) enable agent functionality, while commerce protocols (UCP, ACP) enable transactions. Companies that understand both layers will be positioned to capture value as AI agents become primary transaction initiators.

The Implementation Imperative

Businesses don't need to implement all four protocols immediately, but they need to understand the strategic landscape. For companies with existing Schema.org markup, NLWeb represents the closest on-ramp. For businesses with APIs or internal tools, MCP accessibility should be considered. For organizations evaluating multi-vendor agent workflows, A2A is the protocol to watch. For development teams using AI coding tools, AGENTS.md should be adopted now.

The underlying message is consistent: the agentic web is being built on open standards, not proprietary ones. Companies that understand these standards early will be better positioned as AI agents become a primary way users interact with businesses. The protocols are being established now, the governance is in place, and the agents are already using them. The time for strategic positioning is now.

Source: Search Engine Journal