From grounded to full altitude
Always start small: one team and one product, not the whole organisation at once. Four phases, three gates, and a clear picture of where you stand before you begin.
Four phases, three gates
Each transition has a gate with clear criteria, so nobody scales up before the previous phase works.
- About 2 weeks
Kite Scan
Baseline and interviews. Zones mapped.
- About 6 weeks
First flight
One pilot team. Brief, zones and loop running.
Gate 1
Pilot team chosen, baseline in place - About one quarter
Steady flight
Strengthen the Tail. Roles and metrics in place.
Gate 2
Rework not higher, team wants to go on - Ongoing
Full altitude
More teams on board. Internal Kite coaches.
Gate 3
Zone mix shifting, review load manageable
Durations are indicative and agreed per client. Each gate must be passed before the next phase starts.
What you have at the end of each phase
Kite Scan
A report with the maturity level per part, a baseline of the six metrics, and a proposal for the pilot team and initial zone routing.
First flight
A working Kite Brief template, zone routing for the pilot team, a running Kite Loop with Daily Demo, and a first Human Anchor list.
Steady flight
A stronger Tail, Kite roles filled, a metrics dashboard, and the first recalibration of the zones.
Full altitude
Rollout to more teams, trained internal Line Keepers and Kite coaches, and the method embedded in onboarding.
How high does your team already fly?
The Kite Scan places each part of the kite on one of four levels.
Grounded
AI is used ad hoc, with no shared way of working.
Taking off
AI builds along, but review and verification lag behind.
Steady flight
Zones and the Kite Loop are running and the metrics are in place.
Full altitude
Most work flies in Free and Guided, with a strong Tail and living anchors.
Eight scan questions
These are the questions the Kite Scan starts from, two per part of the kite. Answering them honestly already shows where your line is slack.
| Part | Scan question |
|---|---|
| Frame | Does work start with a testable description of the goal and acceptance criteria? |
| Frame | Are architecture rules captured in a form an agent can read? |
| Sail | What share of the build work do AI agents do today, and who steers them? |
| Sail | Do agents work on several items in parallel, or one at a time next to a developer? |
| Tail | How confident is the team that green tests really mean “works”? |
| Tail | Is every production failure turned into a test or guardrail? |
| Line | Is there a named human who signs off each release? |
| Line | Can juniors explain what they deliver, even when AI built it? |
Find out how high your team already flies.
In about two weeks, a few conversations and a short questionnaire give you a maturity level per part of the kite, a baseline of the six metrics and advice for your first flight.