# JetBrains AI for Teams Is Live: Govern Claude Code, Codex, and Gemini CLI > Source: > Published: 2026-07-11 02:08:29+00:00 Ninety-seven percent of enterprise dev teams are running AI coding tools. Fewer than thirty percent have any governance over them. That gap — between adoption and control — is exactly what [JetBrains announced it is closing on July 7](https://blog.jetbrains.com/blog/2026/07/07/jetbrains-ai-for-teams-and-organizations-from-fragmented-ai-usage-to-coordinated-software-development/). JetBrains AI for Teams and Organizations is not another AI assistant. It is the management layer sitting above the Claude Codes, Codexes, and Gemini CLIs that your developers are already running. ## The Problem Is Not the Tools — It’s the Visibility Engineering leaders in 2026 are stuck asking three questions they cannot answer: who is using which AI tool, what is it actually costing the org, and is the code it generates safe? A single team might be running GitHub Copilot, Claude Code, Cursor, and a homegrown agent simultaneously — each billing differently, touching different parts of the codebase, with no shared context or coordination. That is not a productivity win. That is technical debt waiting to happen. JetBrains AI for Teams does not solve this by taking tools away. It solves it by adding a coordination and visibility layer above whatever AI tools your team already uses. ## What JetBrains Central Actually Is The centerpiece is **JetBrains Central** — a management dashboard for engineering leaders. It provides org-wide visibility into AI tool adoption, governance controls, access management, and model-level policies. Teams can set budgets by team, individual, or policy. Cost attribution tracks exactly where AI spending is going. Model and agent controls let you specify which AI models different developers or teams can access. Crucially, JetBrains Central is vendor-agnostic. It does not require switching to JetBrains IDEs or JetBrains AI. It connects external tools via the [Model Context Protocol (MCP)](https://www.infoworld.com/article/4195002/jetbrains-seeks-to-unify-fragmented-ai-based-software-development-with-governance-suite.html) and external agents via Agent Client Protocol (ACP). ## JetBrains Central CLI: No Developer Revolt Required The smart move here is the **JetBrains Central CLI**. It brings Claude Code, Codex, and Gemini CLI into the org’s governance environment without changing anything about how developers use those tools. Developers keep their preferred setup. The org gets visibility, analytics, and policy enforcement layered transparently on top. This matters because governance tools that force workflow changes get ignored. Central CLI avoids that trap. ## ACP: The Protocol Worth Understanding Now The piece most developers have not encountered yet is the **Agent Client Protocol (ACP)**. [Created by JetBrains and Zed](https://www.jetbrains.com/acp/), ACP is a JSON-RPC 2.0 standard that lets any AI coding agent work in any supporting editor — JetBrains describes it as “LSP for AI agents.” The Language Server Protocol analogy is apt: once editors and agents both speak ACP, you can mix and match freely without vendor lock-in. As of June 2026, fifty-plus agents implement ACP: Claude Code, Gemini CLI, Codex, GitHub Copilot, Goose, and more. JetBrains and Zed have native support; community plugins cover Neovim, Emacs, and VS Code. AWS’s Kiro IDE also supports it. If ACP establishes itself the way LSP did, JetBrains will have driven one of the more consequential developer standards in years. ## Repository Intelligence and Cloud Agents **JetBrains Context** provides agents with cross-repository knowledge so they spend less time exploring the codebase and more time executing tasks. This directly reduces the number of agent “turns” — which reduces cost and improves output quality. **Cloud Agent Runtimes** let agents run long-running tasks autonomously in managed environments while keeping the work visible and shareable across the team. These two pieces together start to look like a complete agentic development infrastructure, not just a governance tool. ## Who Gets This First and When JetBrains AI for Teams is rolling out gradually to business customers through July and August 2026. Individual and non-commercial users are not in scope for now. An Early Access Program has been running with design partners since Q2 2026. No public pricing has been announced. If you are on a business plan, expect capabilities to become available over the next few weeks. If you are evaluating this for your team, [The New Stack’s breakdown](https://thenewstack.io/jetbrains-ai-team-governance/) of the governance architecture is a solid place to start understanding what you are actually buying. ## The Bet JetBrains Is Making JetBrains is not competing with Anthropic or OpenAI or Google on model quality. They are betting that as AI agents proliferate, the coordination and governance layer becomes the valuable real estate — and that a company with 26 years of IDE experience is better positioned to build it than anyone else. The 97% adoption / 30% governance gap is their market. Whether they can capitalize on it before someone else builds the same layer is the real question. But the problem they are solving is genuinely unaddressed, and the approach — vendor-agnostic, protocol-driven, developer-workflow-preserving — is exactly right.