MCP Apps Enterprise Adoption: Who Supports It and Why It Matters¶
MCP Apps launched as the first official MCP extension on January 26, 2026, co-developed by Anthropic and OpenAI. Within weeks, enterprise platforms began shipping support. This page tracks who adopted it, what the enterprise pattern looks like, and why verification becomes the differentiator once rendering is commoditized.
Enterprise adopters¶
- Airia: announced full MCP Apps support February 12, 2026 as the first enterprise AI platform to embed interactive dashboards, forms, and visualizations directly into AI-powered conversations
- Willow: an MCP gateway that brings MCP Apps to enterprises with UI security, auditing, and governance built in
- WorkOS: documents MCP Apps as "the UI layer" of the protocol, the standard way tools return interactive experiences
- CData and others: enterprise MCP roadmaps treat interactive rendering as a baseline deployment requirement for 2026
The pattern¶
The enterprise pattern for MCP Apps is consistent:
- A gateway or platform provides the rendering surface and the security boundary (sandboxed iframes, audited messages, user consent)
- Tools return UI resources via the ui:// scheme
- The host renders the interface inside the conversation
- Governance applies to the interaction layer, not just the data layer
Why verification becomes the differentiator¶
Rendering is commoditizing quickly. Any MCP server can return a UI resource. The enterprise question is no longer "can you render a dashboard" but "can you make the numbers in the dashboard true."
That is the position that stays open: a layer that fetches live records, validates them against pinned metric definitions, reconciles conflicts across systems, and only then renders. Gateways render. CorpusIQ verifies.