CostHawkCostHawk
Private to Your Organization

Team Benchmarks

Rank your engineers across four dimensions of AI adoption — privately. Only your organization can see this data. It never leaves your account.

Get Started FreeLooking for the public leaderboard?

Scoreboard Dashboard

Four board types, time period filters, team filtering, and full breakdowns — all scoped to your org.

CostHawk Team Scoreboard dashboard showing activity rankings, source mix, token breakdown, and leaderboard trend for an organization

Rankings Table

Percentile-scored rankings with per-member metrics. Click any row to inspect.

Ranked leaderboard table showing scores, tokens, sessions, and active days with a selected member detail panel

Member Deep Dive

Every engineer, fully understood

Click any member in the rankings to see their complete profile — source breakdown, token usage, activity trends, and daily patterns.

  • Claude Code vs Codex source breakdown
  • Input, output, and cache token split
  • Activity trend chart with daily granularity
  • GitHub-style daily usage heatmap
Member detail view showing source mix between Claude Code and Codex, token breakdown, activity trend chart, and daily usage heatmap

How It Works

Four Scoreboards. One Dashboard.

Each scoreboard uses percentile-based scoring across multiple metrics. Your team is ranked by normalized performance, not raw numbers — so small teams and large teams are measured fairly.

Activity

Raw Momentum

Who is generating the most real AI usage right now. Tracks total tokens, session count, and active days to surface your most prolific AI users.

Total tokensTracked sessionsActive days

Consistency

Sustained Usage

Who keeps usage steady instead of appearing for one spike and fading. Rewards engineers who build AI into their daily workflow, not just one-off experiments.

Active daysStreak (consecutive days)Active weeks

Cache Efficiency

Savings Signal

Who is creating the strongest cache-reuse and savings signal. Identifies engineers whose usage patterns generate cost savings through efficient prompt and context caching.

Cache read ratioEstimated savingsActive days

Adoption

Tool Breadth

Who is turning Claude Code and Codex into a repeat habit across tools and providers. Measures breadth of adoption, not just depth of any single tool.

Active daysTracked sessionsTool and provider breadth
Privacy Guarantee

This Is Not a Public Leaderboard

Team Benchmarks are an internal tool for your organization. Your data stays in your account, visible only to authorized members.

Organization-scoped

Scoreboards are scoped to your organization. Each team only sees their own members. No cross-org data is ever exposed.

Role-based access control

Only owners, admins, and authorized members can view rankings. Permissions follow your existing org structure.

No external sharing

Scoreboard data never leaves your account. It is not aggregated, anonymized, or shared with other organizations. This is not a public leaderboard.

Built For

The Questions Your Board Is Asking

“Is the team actually using AI?”

The Activity board shows exactly who is using AI tools and how much — by tokens, sessions, and active days.

“Is it sticking, or was it a one-week experiment?”

The Consistency board tracks streaks and sustained usage patterns, separating real adoption from novelty spikes.

“Are we getting value from the spend?”

The Cache Efficiency board identifies engineers whose usage patterns generate cost savings through intelligent caching.

“Who needs training and who is leading?”

The Adoption board measures breadth of tool usage, helping you identify champions and engineers who need support.

Get Started

See your team's AI adoption in under 5 minutes

Set up CostHawk, invite your team, and get your first scoreboard rankings. Free, private, and no credit card required.

Get Started FreeRead the Docs