Connect AI usage to engineering work.

FML ties every agent session to what came of it: the PR it became, the review comments it drew, the incident it fixed.

Get started

Team usage

Last 30 days · sessions, skills, costs, and connected work.

Active engineers

of 42 in last 30d

Sessions

128

across 9 repos

Tool calls

9.4k

read, edit, test, search

Skills used

in shipped sessions

Recent AI sessions

Fix checkout retry path without widening PCI scope

$4.18

@patrick / checkout/main

PR #398 mergedSentry CHK-212 resolved2 review comments

Trace signup latency from PostHog funnel to worker job

$1.42

@gus / growth/experiment-212

PostHog signup_dropoffPR #412 mergedNotion note reused

Rebuild auth migration notes into a reusable skill

$0.88

@rohan / platform/auth

skill created3 engineers installedreview checklist updated

Connected to the tools where work ships

Usage alone is not enough. FML connects sessions to code review, incidents, product analytics, and docs.

GitHub

PRs, reviews, comments, reverts, branches, and files touched by agent sessions.

Sentry

Incidents and production errors tied back to the sessions that fixed them.

PostHog

Funnels and product events connected to work that shipped.

Notion

Docs and decisions reused inside follow-up agent runs.

Questions this answers

The goal is not to rank engineers by prompt count. The goal is to understand whether AI usage is changing real engineering work.

01Who is using AI heavily, lightly, or not at all?

02Which skills and workflows show up in shipped work?

03Which sessions turned into PRs, incident fixes, docs, or rework?