JetBrains AI Assistant Usage Tracking: What the New Admin Analytics Show (and Still Miss)

JetBrains spent most of 2026 building out JetBrains AI Assistant usage tracking for organizations, not just for individual developers. A February release put adoption and credit data in front of admins. A July early-access CLI added spend limits across any agent. Neither one, read alone, tells a manager whether the rollout is actually working.

Read together, they tell a more useful but messier story: three different consoles, each covering part of the picture, with real version gates and scope limits that JetBrains documents plainly if you go looking.

What does JetBrains' new AI Activity and Impact dashboard actually show?

The dashboard reports AI contribution rate, code acceptance rate, per-tool code origin, and feature activity like chat and suggestion volume, filterable by user, group, and date range.

AI contribution rate divides AI-committed lines by total committed lines, giving a single adoption number per team or org. Code acceptance rate tracks what share of suggestions developers actually keep. Code origin breaks the same data down by source: Junie, Claude Agent, Codex, code completion, and next-edit suggestions each get their own line.

Access is role-gated. JetBrains requires a "View AI analytics," "View AI analytics (self)," or "Manage organization" permission to open any of it, and admins can scope those roles at the company or team level.

The version gate that matters most. Contribution rate and code origin are IntelliJ IDEA-only and need IDE 2025.3.0+ with current plugin versions, 2026.1+ for the full metric set. If your team runs PyCharm, WebStorm, Rider, or GoLand, you see adoption and credit numbers, not this deeper code-impact view, until JetBrains extends parity.

What does the dashboard still miss?

It misses uncommitted code, Central CLI activity, JetBrains Air activity, and any tool running outside the JetBrains ecosystem entirely.

JetBrains documents these gaps directly rather than leaving them implied. The dashboard only counts code that reaches a commit, so exploratory sessions that get discarded never show up. Central CLI and Air both report through their own surfaces, not this one.

What you want to know	AI Activity and Impact dashboard	Where it actually lives
AI contribution rate, IntelliJ IDEA	✅ 2026.1+ only	This dashboard
Same metric, PyCharm/WebStorm/Rider/GoLand	❌ Not yet available	Not shipped
Credit consumption and active users	✅ All IDEs	AI Credit Consumption view
Cross-tool spend limits (incl. third-party)	❌	Central CLI (separate console)
Uncommitted or discarded AI sessions	❌	Not tracked anywhere
Standalone Claude Code CLI, Cursor, Copilot	❌	Not tracked by JetBrains at all

A team that only checks this one dashboard will see a clean adoption number and miss three entire categories of activity sitting just outside its scope.

What is JetBrains Central CLI, and why doesn't its data show up here?

Central CLI is a separate early-access product for setting granular AI spend limits across every agent and IDE, including tools JetBrains doesn't make, and it reports through its own console rather than the Activity and Impact dashboard.

JetBrains launched Central CLI in early access on July 8, 2026, aimed at mid-size organizations with heavy API-based usage across multiple third-party tools. It lets admins set token budgets by department, developer, and agent, and it aggregates forecasted cost across all connected AI expenses.

JetBrains is candid about who this is for. The company states plainly that "if your AI usage is not that big and costs aren't yet a pressure point, the economics probably won't work in your favor," and notes that coverage today is limited to three popular terminal agents, with four more still in beta. Personal AI subscriptions and niche setups stay out of scope entirely.

That combination, real spend-control value for large multi-tool orgs, limited agent coverage, and a separate console from the main analytics view, means a team adopting both Central CLI and the Activity dashboard is running two different reporting surfaces for one AI rollout.

How does JetBrains AI Assistant tracking compare to GitHub Copilot and Claude Code?

JetBrains now reports more native usage data than Claude Code ships on its own, but its deepest metric is IDE-gated, and like every native tool, it stops at its own ecosystem's edge.

Metric	JetBrains AI Assistant (Console)	GitHub Copilot	Claude Code	CloudByte AI
Team-level adoption view	✅ Credit/active-user view, all IDEs	✅ Dashboard + API	❌ No team dashboard	✅
Code-impact metric (AI vs. human lines)	🔶 IntelliJ IDEA 2026.1+ only	❌ Not native	❌ Not native	✅ Via commit analytics
Public API for pulling data	🔶 Promised, not broadly shipped	✅ REST API	❌	✅
Cross-tool normalized view	❌ JetBrains-ecosystem only	❌	❌	✅
Tracks tools outside its own product	❌	❌	❌	✅
Ghost seat / never-activated alert	❌ Not documented	🔶 last_activity_at, no alert	❌	✅ Automated

The practical read: JetBrains has closed part of the gap that Windsurf still has and that Claude Code never had natively. What it hasn't closed, and says so itself, is a single view spanning every tool a mixed-IDE team actually uses.

How do you measure JetBrains AI Assistant adoption and ROI across a team?

Combine the credit-consumption and active-user data every IDE reports with your own git commit history, instead of waiting on the IntelliJ-only contribution-rate metric to reach the rest of your toolchain.

A practical sequence that works regardless of which JetBrains IDE a developer runs:

1. Pull active-user and credit-consumption trends monthly. These metrics work across all JetBrains IDEs today, unlike contribution rate.
2. Flag users who repeatedly hit their monthly credit limit. JetBrains surfaces this directly; a cluster of limit-hitters in one team is a capacity-planning signal, not just a cost one.
3. Cross-reference credit spend against commit volume per developer. A developer burning credits with no matching rise in merged PRs is worth a direct check, the same pattern ghost seat detection looks for on any seat-billed AI tool.
4. Treat the IntelliJ-only contribution metric as a leading indicator, not the full picture, if your team is mixed-IDE. A PyCharm-heavy backend squad won't show up in that number at all yet.

If your team also runs Claude Code. The Console can restrict which agents, including Claude Agent, developers are allowed to use inside JetBrains IDEs, but a standalone Claude Code CLI session started in a terminal is invisible to all three JetBrains consoles. BYOK Claude Code setups and native session capture cover that side separately.

Do you need a separate tool if you already have JetBrains' analytics?

Yes, if your team uses more than just JetBrains AI Assistant, which most mixed-tool engineering orgs do.

JetBrains' own roadmap acknowledges the current analytics are a first step: the company says "more advanced metrics and API access" are coming in upcoming releases, with centralized key management, MCP administration, and audit features still in development. That's a reasonable pace for a product that only shipped its organization-wide analytics in February 2026.

For a team running JetBrains AI Assistant next to Claude Code, the gap isn't a quality problem with JetBrains' dashboard. It's that no single console, JetBrains' or anyone else's native one, reports both. CloudByte AI captures Claude Code sessions directly from the developer's own environment and normalizes that alongside other tools, filling the specific hole JetBrains' three consoles leave open. See how the same multi-console gap plays out for GitHub Copilot teams →

If you're comparing options for Claude Code-specific tracking rather than the JetBrains side of a mixed rollout, the full roundup of Claude Code monitoring tools covers ten of them side by side. Pricing for either side of a mixed rollout is on the pricing page, or book a demo to see the per-session dashboard directly.

---