01
Every AI tool, every team, every week.
Larridin reads cost signals from every AI coding tool integration in your stack and reconciles them at the engineer, team, and department level.
Cost Analytics
Every AI dollar.
Tied to the work it produced.
Spend reports tell you who logged in. They don't tell you who shipped. Larridin closes the gap by reconciling every dollar of AI spend against the actual output it produced. By tool, by team, by user, every week.
Why most AI spend reporting fails
Every CFO has a stack of AI invoices.
None of them answer the only question that matters.
Our AI update to the board was a list of subscriptions. That can't happen again.
Per-vendor seat counts. License utilization reports. Cost-per-user dashboards. They all answer the same shallow question: who logged in.
The question the board is asking is different. Is the spend producing work? Is the team that's spending the most actually shipping the most? Are the engineers with Copilot turning out cleaner code than the ones without? Where can we recover budget without losing productivity?
No admin panel answers any of that. Vendor admin panels show usage. Procurement tools show contracts. Neither one sees the work product the spend is supposed to be enabling.
Today · In production
Cost intelligence, fully realized for engineering.
Three integrated views that no admin panel for Cursor, Copilot, Claude Code, or any other tool can produce.
02
Every dollar, matched to the work it produced.
Larridin reads cost signals from every AI coding tool integration in your stack and reconciles them at the engineer, team, and department level.
03
Are heavy AI users producing more, or just spending more?
Filter every quality metric by AI usage level. Compare cost-per-PR, innovation rate, and code turnover by tier. The cost question becomes the productivity question.
Under the hood
Cost signal in. Work signal in. Truth out.
Larridin connects to your AI tool integrations and reads cost signals — license utilization, per-user spend, agent invocations, token consumption — directly from each tool's admin surface. Independently, Scout's behavioral telemetry observes what work is actually happening: PRs merged, code shipped, workflows completed, AI-touched output.
Scout reconciles the two streams at the user and team level. The output is a continuous ledger of cost-to-output, updated automatically as new data lands.
The framework extends
The same framework runs wherever AI runs.
Engineering AI is the most instrumented surface in the enterprise today, which is why cost intelligence is mature there first. The same reconciliation extends to every other department as integrations come online, answering the same shape of question for each one.
Same framework. Cost in, work in, reconciliation out. As integrations across each surface come online, Cost Analytics extends.
Built for engineering leaders
Four views that change how you talk about AI spend.
Team performance · Individual breakdown
Heavy users at $12K in cost shipping 321 PRs. Light users at $300 shipping 114. By name.
See Team Performance · Individual Breakdown →Engineering insights · Cost & ROI
$41K total cost, normalized against PR volume, cycle time, and code quality, by team.
See Engineering Insights · Cost & ROI →Cost view · Recoverable budget callout
$94K/yr recoverable on Copilot seats showing no measurable lift. By name.
See Cost View · Recoverable Budget Callout →Quality · AI usage level filter
Filter every quality metric by AI usage tier. Innovation rate, turnover, bug fixes — all sliced by usage level.
See Quality · AI Usage Level Filter →Scope
Cost Analytics doesn't replace your procurement system.
It does something procurement can't.
Larridin is not a contract management tool. We don't negotiate renewals, route purchase orders, or manage vendor relationships. The procurement stack you have is the procurement stack you keep.
What Scout adds is the layer procurement was never built for: continuous reconciliation between the AI dollars that are already flowing and the work those dollars are actually producing. The decisions you make about contracts, seat counts, and vendor renewals get sharper because the inputs to those decisions are real instead of estimated.