The 4 Key Results of Agile Together

Completion rate, rising velocity, consistency of the events and AI leverage — fixed for everyone, and calculated from your board.

Your Objective is yours. Your Key Results are not: this course uses the same four for
everybody. That surprises people, so here is the reasoning before the definitions.

Why they are fixed

Your Objective is unique — launch a podcast, get fit, ship a feature. The Key Results
here do not measure whether you achieved that. They measure whether you are
executing the method that would get you there. Cadence, follow-through, honest
sizing, real use of the AI.

KR1 — Completion rate

Of everything you committed to in a Sprint, how much reached the Completed
list. Target: 80%.

Below that and you are over-committing at Planning — which is normal in Sprint 1 and
a problem by Sprint 3. Consistently at 100% and you are probably under-committing: a
Sprint that never strains tells you nothing about your capacity.

KR2 — Rising velocity

Cards completed in the last Sprint, compared with the one before. There is no fixed
target: the first Sprint sets your baseline, and the point is the direction of travel.

Velocity is not a productivity score and it is not comparable between people. It is
your own capacity, measured rather than guessed — which is what makes the next Planning
honest.

KR3 — Consistency of the events

Scrum works because the events happen, not because they are scheduled. This KR counts
the events actually documented on your board against the events the cadence calls for —
seven per Sprint at the default settings: one Planning, four Dailies, one Review,
one Retrospective.

You set the cadence once, the way a real team chooses its timebox. Four Dailies is a
one-week Sprint; around nine is two weeks.

KR4 — AI leverage

How many tasks that came from the AI were actually delivered — marked
[AI] on the board and finished. Target: five per Sprint.

This is the KR people misread. It does not reward asking the AI for things; it rewards
finishing what came out of that conversation. A hundred suggestions and nothing
delivered scores zero, and it should.

Reading them

Step 10 of the menu — the weekly OKR Check-in — prints all four with a traffic light,
and hands you a prompt so you can talk them through with your AI. Confidence in the
check-in reads green ≥70%, amber 40–69%, red <40%.

Next

Lesson 8 sets up ProjectAdm — the account, the token, and the board that becomes your project.

Agile Together is a free course. The package — scripts plus a portable Python — is a single download: get the course package. You install nothing.