The nature of your work,
not just your activity.

Work intelligence that classifies what you actually do (innovation, maintenance, glue, tech-debt, firefighting), so you can see your trajectory, your automation risk, and the skills worth growing into.

read-only GitHub access · no training on your data · private by default
problem

Activity metrics were always a proxy. In the AI era, they break.

Productivity looks fine until you ask what kind of work got done, and who is being automated out of a job they still appear busy in.
the visible
Commits, tickets, PRs. These count output, not nature. A commit that wires GPT into a form field looks identical to one that designs a new architecture.
the invisible
Reviews, mentoring, unblocking, design docs, cross-team wiring. The work that makes teams function goes uncredited because nothing counts it.
the consequence
The people whose work is most automatable look productive right up until the moment they do not. Nobody sees the cliff. Not the company, not the employee.
how it works

Three quiet steps. No dashboards to configure.

01

Connect your work

Sign in with GitHub. Read-only, scoped to repositories you choose. Nothing leaves the pipeline and nothing trains a model. More sources come next: Linear, Jira, Slack, calendar.

02

See your work classified

Every commit, PR, and review is classified semantically (innovation, maintenance, glue, tech-debt, firefighting) with a confidence score and a one-sentence rationale you can audit.

03

Grow with clarity

Watch your mix change over weeks. See automation risk on specific kinds of work. Get suggested skill shifts grounded in what you already do well, not a generic career map.

audiences

Built for the person doing the work, and for the team around them.

Two views of the same data. Different incentives. Zero leakage from the personal to the team side.

Trajectory

Your work mix over weeks and months. See whether you're drifting toward maintenance or deepening in new capability, before anyone else notices.

Automation risk

A rolling score grounded in template density, novelty, and cross-module coupling. Specific patterns flagged, not vibes, so you can act on them.

Skill shifts

Career moves rooted in what you already do well, not a generic map. If you review a lot of infra PRs, that's signal, and we tell you where to point it.

principles

Opinions we will not trade away.

These are not tactics. They shape every design decision and every feature we refuse.

01

Employee-first

Your data lands in your view before anyone else's. The personal dashboard is the product; team views are an aggregation on top, never the other way around.

02

No leaderboards. Ever.

No individual rankings visible to managers. No top-performer lists. Teams see distributions, not names, by design, not by policy toggle.

03

Semantic, not metadata

Classification reads the actual work (diffs, messages, PR descriptions) and explains itself. We do not trust Jira labels to describe reality.

04

Credit invisible work

Reviews, mentoring, design docs, unblocking. First-class contributions, not footnotes. The threads that hold teams together deserve to show up in the weave.

roadmap

Starts with engineering. Ends nowhere near it.

Software commits are the densest, most structured signal of knowledge work, which makes engineering the right place to prove the method. The same classification layer maps cleanly onto product, design, analytics, and operations.

now · liveengineering · GitHub
nextproduct · Linear, Jira
laterdesign · Figma
lateranalysts · notebooks, dbt
laterops · calendar, docs

See the shape of your work. Before someone else does.

read-only · private by default · no training on your data
five threads · one signature