What a gate record actually looks like
Programmes talk about gates as if the meeting were the control. It is not. The meeting is an event, and events are forgotten. The control is the record that the event leaves behind.
The AI Change Loop runs four gates: Mobilize, Activate, Go-Live and Handover. At each one somebody decides whether the programme proceeds, proceeds with conditions, or does not proceed. What matters later is whether that decision can be reconstructed by someone who was not in the room.
Five things the record has to hold
The decision itself, stated in a form that can be read back. Who recorded it. The evidence it rested on, referenced rather than summarised from memory. Any condition attached to it. And whether each condition has since closed.
The last two are the ones that go missing. A conditional proceed is the most common outcome at a gate, and the condition is the part that carries the risk. If it is not written down with an owner and a closing test, it behaves as an informal promise, and informal promises do not survive a change of role.
Why one recorder matters
When several people can record a decision, several versions exist. When the person recording is also the person pressing for the outcome, the record bends toward the outcome. A single accountable recorder keeps the log honest and gives the next reader one place to look.
This is the discipline LR-5, the Governance Reallocation Decision Log, exists to enforce. It treats the log as a working instrument rather than an archive, because a log nobody reads at the next gate is only a filing system.
The gate is the meeting. The control is the record.
A test you can run this week
Take the most recent gate decision your programme made. Find the record. Check that it names the recorder, points to the evidence, lists every condition and shows which conditions have closed. If it takes more than a few minutes to assemble, or if the answer depends on who you ask, the decision has already started to decay.
LR-2 then asks the follow up question: at what cadence is that record revisited, and by whom. A record that is accurate on the day it is written and never reviewed is the governance problem in miniature.
The absorption ceiling nobody scores
Concurrent initiatives draw on the same finite capacity, and no single programme can see the total draw. The ceiling is real, it is measurable, and almost nobody measures it.
Shadow workflows are quietly holding your adoption numbers up
Go-live does not eliminate the parallel process. It drives it underground, where it keeps the numbers looking healthy while the transformation quietly fails to land.
Nobody can prove it worked, and that was decided at the start
The inability to demonstrate return is not a reporting failure discovered at the end. It is a design decision taken in the first ninety days, when the value case was written in a form that could never be tested.
Start with the free diagnostic
Ten questions, an indicative readiness band, and one prescribed next step. Returned on screen, no call required.
See what it measures