# MCP Won: How Anthropic's Protocol Became the Plumbing for Every AI Agent Platform

The Model Context Protocol, created by Anthropic, is now an open standard adopted across Zapier, n8n, and Make - letting all three ship native AI agent capabilities on the same underlying plumbing at once.

By Priya Nair (AI Industry Reporter) — published 2026-09-10, updated 2026-09-22
Source: https://aiscoutdaily.com/news/mcp-became-the-standard-for-ai-agents

## TL;DR
MCP (Model Context Protocol), originally created by Anthropic, is now an industry-wide open standard for how AI agents connect to tools and data - and its adoption across Zapier, n8n, and Make is why all three platforms shipped native agent features around the same time. Survey data cited by Zapier puts AI agent usage at 72% of enterprises for tasks like data management and customer support, with n8n's self-hosting model becoming the deciding factor for enterprise engineering teams specifically.

A year ago, "AI agent" mostly meant a chatbot with extra steps. This year, it means something with a defined technical shape - largely because of a protocol most end users have never heard of. MCP, the Model Context Protocol, was created by Anthropic and has quietly become the open standard for how an AI agent connects to tools and data. That standardization is the actual reason Zapier, n8n, and Make all shipped native agent capabilities within the same stretch of 2026, rather than each building incompatible, proprietary versions of the same idea.

## What changed on each platform

- Zapier launched Zapier Agents for autonomous task execution across its 8,000+ app integrations - though it's worth being precise about what this is: Zapier treats AI as a feature layered onto a workflow, not as an architecture. You can call AI APIs, but you can't build a stateful agent with persistent memory and context the way n8n's architecture allows.
- n8n shipped a 2.0 release in January 2026 with native LangChain integration and roughly 70 AI-specific nodes. Its self-hosting model - long a niche feature for compliance-sensitive teams - became a decisive advantage as data sovereignty concerns pushed more enterprise engineering teams toward it specifically.
- Make introduced Maia, an AI assistant that builds automation scenarios directly from a natural-language description rather than requiring manual node-by-node setup.

## The adoption numbers behind the shift

Survey data cited by Zapier puts AI agent usage at 72% of enterprises for tasks like data management and customer support - a level of adoption that's easy to miss if you're only paying attention to consumer-facing AI news. Self-hosted automation adoption specifically surged 55% in 2025, largely enterprises moving away from expensive SaaS tiered pricing toward infrastructure they control directly - a trend that maps closely onto why n8n, the self-hostable option of the three, is described as the breakout winner among enterprise engineering teams specifically, even as Zapier remains the more approachable option for non-technical teams.

## What this means if you're choosing a platform

The practical takeaway: if your team needs genuinely stateful agents with memory and you have any engineering capacity, n8n's architecture and self-hosting option now have a real, standardized reason (MCP) to be compatible with a much wider range of tools than a year ago. If you need something a non-technical team can operate without engineering support, Zapier Agents remain the easier on-ramp - just go in clear-eyed that it's AI-as-a-feature, not a full agent architecture.

## Sources
- [State of agentic AI adoption survey [2026] - Zapier](https://zapier.com/blog/ai-agents-survey/)
- [n8n 2026 Updates: Release Notes, AI Agents & MCP Integration Guide](https://nodesify.com/blog/n8n-workflow-automation-guide-2026)