Anthropic Launches Claude Tags: Slack-Native AI Coworker

Anthropic launched Claude Tag on June 23, 2026, in beta for Claude Team and Enterprise Slack users [9][2]. The feature turns Claude into a persistent Slack team member that can be @mentioned in channels, read approved conversations, use approved tools, and proactively participate in threads to move work forward [1][3]. The launch directly targets Perplexity AI's 'Computer' Slack integration, which had already seen internal enterprise adoption before Anthropic's move [4]. Reception was broad: Andrej Karpathy called it a 'new paradigm,' The Register described it as a 'nosy, always-on' coworker, and critics warned the feature does not resolve AI hallucination risks in production contexts [6][10][11].

Claude Tag pushes Anthropic beyond the chat-window model into persistent, ambient enterprise presence — a distribution play that embeds Claude into the workflow layer where knowledge workers already spend their day. It opens a direct competitive front with Perplexity in the emerging Slack AI coworker category, testing whether enterprise teams will treat AI as a literal colleague rather than a tool they visit on demand.

Anthropic launched Claude Tag on June 23, 2026, in beta for Claude Team and Enterprise customers. Unlike the prior Claude Slack app, which required users to open a DM or invoke a bot explicitly, Claude Tag joins channels as a standing member — it can read approved conversations, use approved tools, offer unsolicited suggestions, and execute delegated tasks when @mentioned [1][2]. Anthropic's support documentation describes it as a Claude that is 'always-on, working like a teammate you'd expect on Slack' [3].

The launch was framed by observers as a direct response to Perplexity AI's 'Computer' Slack integration. SemiAnalysis noted that Perplexity's Computer had already proven 'wildly useful internally' at their firm, and called Claude Tag a 'direct competitor' [4]. Perplexity's own product description of Computer in Slack emphasizes the same value proposition — moving from 'shared context to finished work' inside existing channels [5]. The two products now occupy the same category, competing on model quality, tool integrations, and enterprise trust.

Receipt on launch day was broad but varied in tone. Andrej Karpathy provided the most substantive public endorsement, calling it 'a new paradigm for interacting with Claude that is significantly more inline with all the other human activity org-wide' and noting the substantial under-the-hood engineering required to make it work [6]. ZDNet, The New Stack, Finimize, and Techmeme covered the launch positively [2][7][8][9]. The Register took a cooler view, headlining Claude as a 'nosy, always-on agentic AI coworker' [10]. On social media, at least one critic warned that 'shiny Slack tags won't hide the hallucinations in your diff' [11], and a self-described former Claude enthusiast said the feature had not won them back [12].

Alongside Claude Tag, Anthropic also previewed 'Claude Code in Slack' in partnership with Salesforce [13], and maintains a separate 'Claude Cowork' product page [14]. Whether these represent a unified product suite, different tiers, or overlapping rollouts is not yet clearly documented. The broader pattern — ambient Claude across enterprise communication and development tools — aligns with Karpathy's framing of a long engineering investment to make AI genuinely inline with how organizations work [6].