Claude in Slack: what Claude Tag is and how to adopt it in your company

In June 2026, Anthropic brought Claude into Slack in a new way. Claude Tag turns the assistant into a teammate who lives in the channel, follows the conversation context and performs tasks alongside people. And because Slack is owned by Salesforce, this development fits directly into the ecosystem your company already uses.

For those who operate with Salesforce, it is worth separating the two layers of artificial intelligence that now coexist. Agentforce is the outward-facing AI: agents that serve customers and act on the CRM. Claude Tag is the inward-facing AI: a colleague that helps the team work better within Slack. One takes care of the customer relationship; the other takes care of operational productivity.

This guide directly explains what Claude Tag is, what changes compared with the old Claude in Slack, how it works and how to adopt it with governance. WeeNow is an authorized Salesforce channel and licenses Slack, so it follows this adoption from data through security.

What is Claude Tag?

Claude Tag is Anthropic's new way of placing Claude in a Slack channel as a shared AI teammate, available to the entire team rather than as an individual assistant. It entered public beta on 23 June 2026 for Claude Enterprise and Team customers.

Claude Tag is an instance of Claude that lives inside a Slack channel, visible and usable by all members. It can understand the conversation context, execute tasks with the tools it receives and reply in the thread itself.

Under the hood, it runs on Claude Opus 4.8, Anthropic's most capable model. The difference is not only its power, but its format: instead of each person talking to their own assistant in a separate window, the whole team works with the same Claude, in the same place where operations already happen.

What changes compared with the old Claude in Slack?

The main change is one of mindset: the former app was individual, while Claude Tag is shared by channel. Previously, each person had their own private conversation with Claude within Slack. Now there is one instance per channel, which everyone can see and call on, with the context of what the team is discussing there.

There is also a deadline to note. According to Anthropic, Claude in Slack will automatically migrate to the Claude Tag experience on August 3, 2026. Administrators have a 30-day window, open since the June 23 launch, to opt in and configure the transition at their own pace instead of inheriting the automatic migration. Those already using Claude in Slack need to decide on the transition; those who do not yet use it have a good opportunity to start correctly.

Another new feature is ambient mode. In addition to responding when called, Claude Tag can act independently within defined limits: update the team on a topic, flag something relevant that appeared in the organisation, and revisit threads or tasks that were left behind. It stops being merely reactive and starts following the flow.

How does Claude Tag work in Slack?

In practice, an administrator connects Claude Tag to the workspace, defines which tools and data it can access, which channels it can operate in and the spending limit. From there, anyone in that channel can tag @Claude with a request and it executes it.

The cycle is easy to describe. Someone asks it to write a text, retrieve sales numbers, or run an analysis. Claude Tag interprets the request, breaks it into stages, executes each step with the tools it has been given, and returns the result in the thread, where the team follows and continues the conversation. Because the instance is unique per channel, history and context are available to everyone rather than being trapped on one person's screen.

This design makes Claude Tag closer to a coworker than a chatbot. It operates where decisions happen, with controlled access to what the company allows.

Why does this matter for Salesforce users?

Because Slack is owned by Salesforce, and Salesforce and Anthropic have formed a partnership that makes Claude a preferred model within the ecosystem, from Slack to Agentforce 360. The same AI Salesforce is bringing to the CRM also comes to live in the team's collaboration tool.

This approach has a clear focus on trust. The partnership targets regulated sectors such as financial services, healthcare and life sciences, and provides trust boundaries so customer data is not used to train the model. For enterprise buyers, this changes the conversation: it is not about trying an isolated tool, but about adopting AI in an environment that already follows company rules.

If your operation runs on Salesforce and Slack, Claude Tag is not an external component. It belongs in the same home.

Agentforce and Claude Tag: the two AI layers of your operation

It is worth seeing both as complementary layers: Agentforce is the outward-facing AI and Claude Tag is the inward-facing AI. One works with customers; the other works with the team.

O Agentforce is Salesforce's autonomous agent platform. These agents serve, decide and act on CRM data in channels such as chat, voice and WhatsApp, always within company limits. It is the layer that interacts with those on the outside: the customer.

Claude Tag is the inside layer. It helps team members be more productive in everyday work within Slack: drafting, analyzing, finding information and moving tasks forward. It does not speak to end customers; it serves the people operating the business.

Together, they create an end-to-end agentic operation, with AI in service and AI in internal productivity. And as a Salesforce-authorised channel, WeeNow covers both fronts: from Agentforce adoption in the CRM to Claude Tag adoption in Slack.

Is Claude Tag secure? Governance, data and LGPD

It is secure when adoption is governed. Control is in the administrator’s hands: they define which tools and data Claude Tag can access, which channels it works in and the spending limit. The Salesforce–Anthropic partnership also provides trust boundaries so data does not feed model training.

The precautions are the same as for any autonomous AI. Granting access that is too broad increases the risk surface, so the path is minimum permission by channel and use case. When Claude Tag handles personal data in Slack, LGPD applies, requiring a legal basis, minimisation, and traceability of what the AI accessed and why.

This is exactly where a well-run implementation makes a difference. Turning on Claude Tag is quick; making it secure requires designing permissions, channels and policies before scaling.

How to adopt Claude Tag in your company

The path is to start with the right foundation and grow with control:

• Ensure the appropriate license: Claude Enterprise or Team, and Slack for your operation.
• Connect Claude Tag to the workspace and choose the channels where it can operate.
• Define tools, data, and spending limits by channel, with minimum permission.
• Establish governance: use policies, personal-data handling and traceability.
• Choose the first highest-leverage use cases and measure the gains before scaling.

WeeNow joins this journey as a Salesforce-authorised channel. It licenses Slack, implements Claude Tag, and designs governance so AI works safely within your operation instead of becoming an uncontrolled standalone pilot. The idea is simple: make inward-facing AI work, from the first channel to the outcome.

Frequently asked questions about Claude Tag

What is Claude Tag?

Claude Tag is Anthropic's way of putting Claude inside a Slack channel as an AI teammate shared by the whole team. It understands conversation context, carries out tasks using the tools it is given and replies in the thread itself, rather than being an individual assistant.

What is the difference between Claude Tag and the old Claude in Slack?

Claude in Slack was individual: a private conversation per person. Claude Tag is a single instance per channel, shared and visible to everyone, with team context and an ambient mode that acts independently within defined limits.

When does Claude in Slack become Claude Tag?

According to Anthropic, automatic migration takes place on 3 August 2026. Administrators have a 30-day window, open since the 23 June launch, to opt in and configure the transition rather than inheriting the automatic change.

Does Claude Tag have anything to do with Salesforce?

Yes. Slack belongs to Salesforce, and Salesforce and Anthropic have a partnership that makes Claude a preferred model in the ecosystem, from Slack to Agentforce 360, focusing on regulated sectors and trust boundaries for data.

Is Claude Tag the same as Agentforce?

No. Agentforce is the outward-facing layer, with agents that serve customers and act on the CRM. Claude Tag is the inward-facing layer, an AI colleague that helps the team in Slack. The two complement each other.

Is Claude Tag secure for company data?

It is secure when adopted with governance. The administrator controls tools, data, channels and limits, and the Salesforce–Anthropic partnership provides trust boundaries so data is not used for training. When personal data is involved, the LGPD applies, with least privilege and traceability.

Sources

• Anthropic, Introducing Claude Tag: https://www.anthropic.com/news/introducing-claude-tag
• Anthropic, expanded partnership with Salesforce: https://www.anthropic.com/news/salesforce-anthropic-expanded-partnership
• Salesforce, business context, and AI actions in Claude through Slack and Agentforce 360: https://www.salesforce.com/news/stories/salesforce-anthropic-trusted-context-ai-actions-on-claude/
• Claude Help Center, get started with Claude in Slack (including migration to Claude Tag): https://support.claude.com/en/articles/11506255-get-started-with-claude-in-slack
• Slack Marketplace, Claude: https://slack.com/marketplace/A08SF47R6P4-claude