Anthropic expands Claude through MCP Apps, embedding interactive tool access directly in the chat to draft, format, and manipulate content in real time within apps like Slack, Canva, and Asana.

Nine integrations launch simultaneously, including Asana, Figma, Canva, Slack, Amplitude, Box, Clay, Hex, and monday.com, enabling live, in-chat collaboration across project management, design, analytics, file access, and more.

The rollout targets enterprises and paid tiers (Claude Pro, Team, and Enterprise) with Salesforce integration anticipated soon as part of the roadmap.

The integrations aim to streamline workflows and automate tasks for both enterprises and consumers, while raising considerations about data consent, security, and AI roles in sensitive environments like healthcare.

Security and guardrails are foregrounded, including consent prompts, IT admin controls over MCP servers, and ongoing safety measures to mitigate risks of real-world agent actions.

Industry implications suggest AI assistants becoming central work hubs, demanding robust permissions, governance, and security frameworks as connectors proliferate.

The broader goal is seamless, ethical, and scalable agentic interactions across business tools and advertising ecosystems.

The piece appears within a sponsor-rich format, highlighting sponsor posts and podcast roundups alongside the tech news.

Claude’s enterprise-driven approach positions it against OpenAI, prioritizing security, reliability, and embedding workflows over consumer-focused features.

Article includes source attribution and keywords related to enterprise AI and the MCP protocol.

This release follows a January safety framework addition to Claude’s reasoning to meet enterprise compliance requirements.

Clear permission scopes, auditable execution logs, and smooth Cowork integration are emphasized to build enterprise trust in agentic actions inside Claude.