Is Claude Tag how we will work with AI tools in the future?

When the AI tool layer disappears entirely, all that's left is a co-worker.

How it works: you invoke a Claude Agent directly from a chat conversation, in Slack for now, by typing @Claude.

What it changes: previously, using Claude meant opening the mobile app, desktop software or web interface. Now that interface disappears entirely. You simply tag Claude in the same place you would speak to a colleague.

What remains is just Claude: present in the conversation, working within its defined role and permissions. Essentially, a co-worker.

Once the complexity of the AI tooling is abstracted, all you see is a name and work product. This might make it simpler to judge work quality without the AI hype distractors.

Note: we have been able to tag Claude and other agents in chats for some time. This is more sophisticated: each agent has its persistent memory and context, its own credentials, work scope, collaborates across teams etc.

It can even jump into conversations to update the team or flag issues. It's more than a simple AI bot. It is also more than just one AI bot (you might have a Finance, an Engineering etc).