Anthropic has flipped a key switch for workplace AI: you can now use Slack, Asana, and Figma directly inside a Claude conversation. Instead of hopping between tabs, Claude can act on your behalf within these apps, show live previews of the work it is about to perform, and let you confirm or edit before anything is pushed to production.
Claude now executes app actions, not just fetching data
Beyond earlier “connectors” that only fetched data, Claude now performs interactive actions from within the chat, mirroring what you would do in the native apps. Built on Anthropic’s open-sourced Model Context Protocol, the system can pull relevant context from authorized accounts, propose a change, and render an in-chat preview that you approve or adjust. At launch, interactive support extends to Slack, Asana, Figma, Amplitude, Box, Canva, Clay, Hex, and monday.com, with Salesforce on the way.
- Claude now executes app actions, not just fetching data
- How to enable interactive Slack, Asana, and Figma in Claude
- What you can do in Slack with Claude’s interactive actions
- What you can do in Asana with Claude’s guided workflows
- What you can do in Figma using Claude’s design actions
- Why building workflows inside chat reduces context switching
- Security, permissions, and governance for interactive actions
- Availability for Claude Pro, Max, Team, and Enterprise plans
- Quick starter prompts to try in Slack, Asana, and Figma

How to enable interactive Slack, Asana, and Figma in Claude
Open Claude on web or desktop and visit the app directory, then connect Slack, Asana, and Figma marked “interactive.” You will be prompted to grant OAuth permissions, which govern what Claude can read and write. Once linked, you simply ask in natural language—Claude fetches the necessary context, proposes the action, and displays a preview panel. You can edit fields inline or choose to open the item in the native app before finalizing.
What you can do in Slack with Claude’s interactive actions
Claude can summarize long threads, draft replies in your tone, and post to the right channel with the right mentions. It can create new channels for projects, set channel topics, and file important messages to knowledge bases via connected tools. Because actions run through your authorized Slack workspace, Claude respects your org’s permissions and will only act where your account is allowed.
What you can do in Asana with Claude’s guided workflows
Give Claude a project brief and it can spin up an Asana project, define sections, generate tasks, assign owners, and set due dates based on your team’s working norms pulled from the workspace. You will see an in-chat preview of the project board and tasks before creation. You can also ask for reporting snapshots—like overdue tasks by assignee—or request bulk edits, with every change previewed first.
What you can do in Figma using Claude’s design actions
For design teams, Claude can create new Figma files, scaffold frames, insert copy, apply components from your libraries, and generate variant names that match your naming conventions. It can also queue up batch layer renames, color token swaps, or text updates for design reviews. A visual preview appears in the chat so you can tweak details before Claude commits changes to the file.

Why building workflows inside chat reduces context switching
Context switching is the tax most teams pay without noticing. Okta’s Businesses at Work report has shown large enterprises commonly run more than 200 apps, and Asana’s Anatomy of Work studies have documented the toll of constant toggling. Research by UC Irvine’s Gloria Mark has also linked interruptions to meaningful losses in focus. Putting core workflows inside a single AI interface promises fewer switches, faster decisions, and cleaner handoffs—especially when the system can read current context and propose the next best action.
It also signals a broader platform play. Competitors are chasing a similar “primary interface” vision—Microsoft via Copilot with Graph connectors, Google via Gemini in Workspace, and OpenAI with agent-style actions. Anthropic’s approach leans on a transparent protocol and human-in-the-loop previews, which may resonate with teams wary of fully autonomous agents.
Security, permissions, and governance for interactive actions
Interactive actions run under the permissions you grant during setup, adhering to least-privilege scopes defined by each app. Enterprises can restrict which integrations are available and which data Claude can access. The preview-first flow means you stay in control: Claude shows exactly what will be created or changed before anything is written back, keeping a clear audit trail of human approvals.
Availability for Claude Pro, Max, Team, and Enterprise plans
The interactive app actions for Slack, Asana, and Figma are available on web and desktop for Claude Pro, Max, Team, and Enterprise plans. Apps marked “interactive” in the directory support in-chat previews and approvals out of the box. Anthropic’s Cowork research preview—aimed at orchestrating multi-step, agentic workflows—is also now accessible for Team and Enterprise customers in addition to Pro and Max.
Quick starter prompts to try in Slack, Asana, and Figma
- “Draft a two-week sprint plan and create an Asana project with tasks assigned to our mobile team,” then review the preview before launch.
- “Summarize the last 24 hours in #customer-feedback and post three takeaways with action items,” to test Slack.
- “Create a responsive hero section using our design system and insert placeholder copy,” then adjust the frame details in the preview.
The bottom line: Claude just moved from helpful narrator to hands-on teammate inside your core tools—without leaving the chat.
