SAN FRANCISCO, January 26, 2026, 11:33 PST
- Anthropic announces Claude now supports opening interactive versions of workplace tools such as Slack, Asana, Canva, and Figma directly within the chat.
- The rollout uses MCP Apps, a new extension of the Model Context Protocol, and wider client support is anticipated
- The change strengthens the connection between AI assistants and actual workplace systems, sparking concerns about control and security
Anthropic launched interactive workplace tools within its Claude service on Monday, enabling users to draft Slack messages, update Asana timelines, and adjust Canva decks—all without leaving the chat interface. The initial lineup features Slack, Asana, Canva, Figma, Box, Clay, monday.com, Amplitude, and Hex, accessible on web and desktop for paid subscribers. Salesforce’s Agentforce 360 integration is set to arrive soon. 1
This shift is significant as AI companies aim to transform assistants from mere discussion tools into actual work hubs. Moving beyond copy-paste summaries to enabling direct edits marks a major leap, one that catches attention when assistants start handling real messages, tasks, and files.
MCP Apps, the first official extension to the Model Context Protocol (MCP), is driving this change. MCP is an open standard designed to link AI products with external tools and data. According to MCP’s core maintainers, MCP Apps enable tools to deliver interactive user interfaces — like dashboards, forms, and visualizations — inside a sandboxed iframe, a secure web frame embedded within the conversation. Claude already supports it, and ChatGPT is set to add support this week. Anthropic’s David Soria Parra said, “I am excited about the possibilities that MCP Apps opens up,” while OpenAI’s Nick Cooper added, “We’re proud to support this new open standard and look forward to seeing what developers build with it.” 2
TechCrunch reported that these embedded apps connect to a logged-in service instance, letting Claude send messages or access cloud files based on the permissions users provide. 3
The Verge noted that while some connectors were already available, they primarily spit out text users had to manually copy into the target app. Anthropic’s new method aims to change that by making tools “open as interactive apps right inside of chat,” enabling users to “see, explore, and refine results visually, not just read about them.” 4
It’s not so much about getting a clever answer as it is about streamlining the process: write, preview, tweak, and send — all within the same thread. For teams buried under countless tabs, that’s the real selling point.
This also highlights the battle over standards. With MCP Apps running on different clients, developers get to create a single interactive tool and deploy it across multiple assistants, rather than having to redo the same integration for every platform.
But granting deeper access is a double-edged sword. A flawed draft remains flawed until someone actually clicks “post.” Companies must weigh how much control and access they’re comfortable giving, particularly with tools handling customer data and internal conversations.
Anthropic is banking on interactive panels inside chat catching on quickly, much like file previews and inline docs did earlier. The real challenge comes after the demo—navigating permissions, policies, and whether companies actually allow these connectors to roll out.
