AI Pulse · Frontier Labs Watch

Slack just made it a lot easier to watch an AI agent do real work before it ships anything. The new feature is called Slack Code. You tag Claude, ChatGPT, Devin, or Copilot in a channel and it spins up a shared space where the whole team can see the conversation, the plan, the live code diffs, and a preview, all before any of it goes live. It’s free on every Slack plan, though access to the actual AI agents is sold separately, according to the announcement.

Here’s why this matters even if you don’t have a dev team touching code daily. Right now, most AI coding work happens in a black box. Someone runs a prompt, gets a result, and the rest of the team has to trust it worked. Slack Code turns that into something visible. A manager, a product owner, or even a non-technical founder can watch the plan form and the diffs appear in real time, in the same tool they already use for everything else. That’s a real change in how AI work gets reviewed and trusted inside a company, not just a coding convenience.

The bigger story underneath this is distribution. Anthropic also previewed Claude Code running natively inside Slack, which puts it in a default position across Salesforce’s massive enterprise customer base. That’s not a benchmark win. It’s a placement win, and those tend to matter more for adoption than raw model quality. If your team already lives in Slack, you’re going to run into these agents whether you sought them out or not.

This kind of visibility into what an AI agent is actually doing, and who’s watching it happen, is exactly the kind of thing we build into an AI command centre. Worth keeping an eye on as more vendors fight for the same “default agent inside your workflow” spot.