Anthropic is rolling out Claude Tag in research preview, an AI assistant that lives inside Slack and operates continuously rather than waiting for a prompt. The feature, accessible to customers on Claude Enterprise and Claude Team plans, lets users tag @Claude in chats to request insights or delegate work.
How it works
This tool builds on existing integrations where users can message @Claude directly or tag it in channels for immediate assistance. Previously, Claude Code in Slack could route coding requests from a channel mention into a full web-based session, returning updates to the thread. Claude Tag adds persistent context and memory that earlier versions could not maintain.
According to Anthropic, the system learns more about the work as it follows along with the channel. It can also gather facts from other parts of the organisation if granted permission to read additional channels.
Every member of a specific Slack channel accesses a single Claude identity. This means anyone can view the assistant’s progress and resume a conversation where the previous user stopped. Administrators control which tools, data, and channels the assistant can reach. Each identity stays scoped to the channels defined by admins, preventing a Claude instance configured for legal work from storing memories in the engineering channel.
When assigned a task, the assistant breaks the work into stages and executes them using available tools. It posts created content in a Slack thread. An ambient mode allows the assistant to proactively join chats to update the team, highlight items from across the organisation, and follow up on neglected threads or tasks.
Anthropic describes the experience as working with a real colleague who produces work in public view with far greater context and understanding than before.
Context is becoming essential for enterprise deployments. Microsoft offers Graph via Copilot and Work IQ. Snowflake and Databricks position their platforms as back-end support containing tacit organisational knowledge that agents can access. Glean is building an intelligence layer that understands company context and sits between the model and enterprise data.
What it means
Teams gain a persistent collaborator that remembers past discussions and proactively manages workflow without needing constant manual intervention.
