The Agent Impact Leaderboard and the Business Impact & ROI Dashboard are live in preview inside GitKraken Insights today. We built them because the questions engineering leaders are getting asked about AI shifted faster than the tools to answer them.

Here’s what shipped and how to get access.

The Agent Impact Leaderboard

The Leaderboard ranks every AI tool and model your team uses across four dimensions: cost, output, rework, and trend. It also identifies your power users and shows how developers at every adoption tier are performing.

Inside it, you can answer questions like: Which tool is winning for new feature development versus code review? What’s our cost per merged PR by tool over the last 90 days? Which developers are Power Users, which are Emerging, and which are still finding their footing? The adoption-tier breakdown (Power User, Regular, Explorer, Emerging, No AI Data) is the cut most teams have never had visibility into, and it’s where the most actionable signal lives.

The Leaderboard is the dashboard you open before a tool renewal conversation. It’s also the dashboard you open when a developer or team lead asks why one tool is being prioritized over another. The data is on the screen.

The Team View

The Leaderboard also gives you a team-level cut. Repo AI readiness, agent adoption tier, speed, and output score for every team in your org, side by side. Use it to see which teams are set up for AI and translating that into output, and which ones might need some help.

The Business Impact & ROI Dashboard

The ROI Dashboard translates engineering output into the language a CFO or board accepts. Productivity delta. Hours returned per developer per week. Dollar value of capacity created. AI spend by tier, with cost per PR alongside.

The math is visible. When the dashboard shows 84 percent productivity gain and $153,000 of weekly capacity created across a 76-developer team, you can see the inputs that produced the number: 1.84x speedup, 34 hours per developer per week, $60 per hour loaded cost. The formulas are inline. Show the math is the design principle, because any ROI number that doesn’t expose its inputs is one good question away from collapse.

This is the dashboard you open when finance, the board, or the CEO asks what the AI investment is returning.

Why now

The AI tooling conversation inside most engineering organizations shifted in the last two quarters. The adoption question is settled. Leadership is no longer asking whether developers are using AI. They’re asking what it costs, what it returns, and which tools are worth renewing.

Most engineering leaders don’t have a good answer yet. The vendor dashboards track seats and usage. The internal dashboards, if they exist, took months to build and break when the person who built them leaves. The result is that engineering leaders are heading into Q3 renewals, board reviews, and headcount conversations with a gut feeling where a number should be.

That’s the gap these dashboards are built to close.

How to get access

Request access at gitkraken.com/insights. Setup takes under a day.