
Tech • IA • Crypto
Anthropic has launched Claude Tag, a Slack-based collaborative AI tool that lets teams interact with a shared Claude instance to automate work, access data, and proactively manage tasks.
Claude Tag integrates directly into Slack, enabling teams to summon a shared AI by tagging “@Claude” in channels. Unlike traditional one-on-one chatbot use, a single Claude instance operates within each channel, allowing multiple users to collaborate with it in real time. Conversations persist, meaning team members can build on prior interactions without restarting context.
Anthropic reports that its own product teams already rely heavily on the system, with 65% of their code reportedly generated using an internal version of Claude Tag. This suggests a shift toward AI-assisted development as a default workflow rather than a supplementary tool.
Claude Tag can connect to external services such as Gmail, CRMs, databases, and productivity tools, pulling data directly into Slack conversations. Through integrations and middleware like Zapier, it can access thousands of additional apps, turning Slack into a central command layer for querying and acting on distributed business data.
Users can assign tasks to Claude, which then executes them in stages and responds within Slack threads. The system supports asynchronous workflows, allowing teams to delegate work and return later with results. Anthropic also संकेत that Claude can handle longer-running processes, potentially working on tasks over hours or days.
A key feature is Claude’s ability to act proactively. In what Anthropic describes as an “ambient” mode, the AI monitors channels and connected tools, flags important updates, and alerts relevant team members. For example, it can detect repeated customer complaints in email and automatically notify both support and engineering teams.
Claude builds contextual understanding over time within each Slack channel, reducing the need for repeated instructions. It can also summarize long threads, highlight missed updates, and identify urgent issues. This persistent memory is scoped to specific channels to avoid cross-team data leakage.
Access to data, tools, and channels is tightly controlled by administrators. Organizations can define which connectors, APIs, and repositories Claude can use, as well as set token usage limits. Data and “memories” remain isolated within designated environments, ensuring separation between departments such as sales and engineering.
The feature is currently in beta for Claude Team and Enterprise users. Anthropic संकेत plans to expand Claude Tag beyond Slack into other workplace platforms, positioning it as a ubiquitous interface for interacting with AI across tools like email, CRM systems, and collaboration software.
Claude Tag signals a shift toward shared, proactive AI embedded directly in team workflows, with Slack as the first step toward broader workplace integration.