Skip to content

Anthropic Brings Data Analytics Agent Claude Tag to Slack for Company-Wide Self-Service Queries

Bottom line: Anthropic is extending its governance-based data analytics approach from Claude Code to Claude Tag in Slack, enabling company-wide self-service access to vetted data answers.

Anthropic has introduced Claude Tag (Public Beta), a data analytics agent for Slack that allows employees without an analytics background to ask questions about company data directly in Slack. The answers are based on the same governance-vetted definitions used by the internal analyst team.

In a previous blog post, Anthropic described how the company brought Claude to around 95 percent accuracy on data analytics questions using three core building blocks: a governance-vetted semantic layer, a collection of skill files encoding internal analytics conventions, and an evaluation suite for measuring answer quality. This approach was initially designed for Claude Code, the primary development environment for data scientists and data engineers at Anthropic.

The new post describes how this foundation is being extended to the rest of the company – via Claude Tag, which is publicly available in beta and serves as the basis for a data analytics agent in Slack. This allows employees outside the data team to also ask data-related questions directly in Slack and receive answers based on the same vetted definitions used by internal analysts.

Anthropic emphasizes that an agent’s technical accuracy and its successful rollout across the entire company are two distinct challenges. The accuracy recommendations already described in the previous post remain valid and are not repeated here. Instead, this post summarizes five key insights from a year of practical experience rolling out a data analytics agent in Slack – with a focus on distribution, permissions management, data freshness, and observability.

For teams looking to deploy Claude or comparable agents for self-service analytics, the post thus offers a firsthand account of organizational and operational aspects that go beyond pure model accuracy – such as how access rights to sensitive data are handled in a broadly accessible chat tool, or how it is ensured that answers are based on current data.


Source: claude.com · Published August 12, 2026
Lumi AI News — AI-assisted curation in accordance with Art. 50 EU AI Act. Paraphrasing and classification by Lumi News Pipeline v1.8.3.

Share on: