Download PDF ↓

Engineering progress,
clearly understood.

Maestro turns code, reviews, and conversations into clear summaries of what your teams accomplish.

  1. Understand progress
  2. Support every team
  3. Recognize contribution
Maestro01 / 11
The question

What did engineering get done this quarter?

Today’s answer

  • 2,140pull requests mergedA renamed variable and a rebuilt billing flow each count once.
  • 860tickets closedA ticket can hold an hour of work or a month. The count can’t tell.
  • 3.4Mlines changedGenerated code and lockfiles move this more than real work does.

All true. None of it answers the question.

Counts measure how much happened, not what changed or why it mattered.

Maestro02 / 11
Core metrics

Three core metrics.

Each is read from the work itself, then rolled up for people, teams, and the whole organization.

  • Code Impact

    The scope, complexity, and effect of every change a person ships.

    Every merged pull request · 0–5
  • Review Impact

    The substance of the feedback a person gives their teammates.

    Every review · 0–5
  • Review Overhead

    The review effort it takes from peers to get a person’s work over the finish line.

    Review points per merged pull request
Maestro03 / 11
Code Impact

Introducing Code Impact

Code Impact measures the significance of a change.

  1. Read every change

    Maestro’s AI reads the code in every merged pull request, not just its size.

  2. Understand what it does

    It works out what the change does and what kind of work it is.

  3. Score its significance

    It weighs scope, complexity, and effect on the system, then scores the change from 0 to 5. Most work lands at 1–3.

Pull requestAdd workspace check to bulk export
+214 −58 · 7 files
 export async function runExport(job) {-  const rows = await loadRows(job.query)+  assertWorkspaceScope(job.actor, job.workspace)+  const rows = await loadRows(job.query, job.workspace)   return writeArchive(rows, job.format) }
SummaryBug

Adds a workspace permission check to bulk export and applies it across the API, the shared exporter, and scheduled jobs.

Code Impact: 3Significant
  1. 0
  2. 1
  3. 2
  4. 3
  5. 4
  6. 5
Maestro04 / 11
The scale

Six levels, from negligible to transformative.

A 3 is a different kind of change, not three times better.

  1. 0NegligibleFormatting, comments, renames
  2. 1MinorA one-line fix or copy change
  3. 2IncrementalBetter error handling in one component
  4. 3SignificantA feature that spans several modules
  5. 4High ImpactReworking a core service other teams rely on
  6. 5TransformativeA major architectural migration. Exceptionally uncommon.
Combining scores

Ten small fixes don’t add up to one transformative change. When scores combine across pull requests, people, and teams, bigger changes carry more weight, so combined scores can exceed 5.

Maestro05 / 11
The organization
Company · 40 engineersLast 16 weeks

Q3 so farCode Impact is up 63% since AI coding agents rolled out, led by Canvas Grove and platform reliability.

Needs attentionPermissions remain staged while Security reviews the new authorization boundary.

AI coding agents rolled out
Code ImpactReview Impact
Jun 29Event retries bounded and traced
Sep 7Canvas Grove focus patterns shipped
Jun 8Sep 21
Code ImpactReview Impact

What advanced, and which way it’s heading.

A written account of the whole organization’s progress, pinned to the trend behind it.

Maestro06 / 11
The work mix
Code Impact by categoryLast 16 weeks
Jun 8Sep 21
FeaturesBugsKeeping the Lights On

See what kind of work drove the change.

Code Impact by category, week by week. After AI coding agents rolled out, feature work accounted for most of the increase.

Maestro07 / 11
The teams

See how each team moved the work forward.

Each team’s progress in a few words, with average Code Impact per engineer so team size doesn’t skew the picture.

Data9 engineers
1.6Avg Code Impact / Eng

Validated event schemas before ingestion, documented a replay-safe recovery contract, and cut nightly pipeline reruns.

Frontend14 engineers
1.5Avg Code Impact / Eng

Standardized keyboard interactions across Canvas Grove and caught focus gaps before they reached product surfaces.

Platform17 engineers
1.6Avg Code Impact / Eng

Bounded event retries, added end-to-end delivery tracing, and moved scheduled exports onto the shared exporter.

Weekly Code Impact per engineer · 12-week average and 16-week trend · dashed line marks the AI coding agent rollout

Maestro08 / 11
The person
Frontend / Tessa R. · Q3
Tessa R.Engineer · Frontend
Last 12 weeks

Tessa R. is improving keyboard interactions in Canvas Grove and catching focus gaps through review.

Recent work
  • Review: nested-dialog focus trapCaught focus escaping on close and proposed a shared restore helper.
    Review Impact3Significant
  • Roving focus for menusArrow-key navigation across every menu component.
    Code Impact2Incremental
  • Shared keyboard patternsOne interaction model for Canvas Grove dialogs, menus, and tabs.
    Code Impact3Significant
Illustrative product view · Fictional data

Recognize the contribution behind the progress.

A written account of their contribution, including reviews, read alongside role, context, and your judgment.

Maestro09 / 11
The workflow

Built on the work already happening.

No new status reports. Maestro reads the systems your teams already use.

GitHubGitLabSlackJiraLinear
  • Your data never trains AI.A contractual commitment.
  • SOC 2 Type II certified.Regular third-party audits.
  • Access stays controlled.SSO and granular permissions.
Maestro10 / 11
The walkthrough

Know what’s happening.
Lead with clarity.

See the progress, support every team, and recognize the people behind it.

Book a walkthrough →

Bring one team and one quarter. We’ll show you their work in Maestro.

Download PDF ↓