AI is being used.But is the work actually getting better?
Organizations measure usage. In a workflow you choose, we field-test where AI actually helps, where human judgment drops out, and what needs to change.
the decision?
Which of these sentences have you said?
Read eight signals along four tension lines. Pick a sentence to see what may sit underneath it and where a leader can start.
Resistance
Resistance is often not disobedience but a signal. Several causes may sit underneath: poor task fit, skill, trust, role concerns, or the system. Ask before you mandate.
Resistance shows up in five behaviors
The eight points above are what a leader feels. What actually shows up in the field is five behaviors.
And each one produces a response in the leader, too. The diagnosis belongs to the moment, not the person — the same person may show a different pattern next month.
See your approach in four minutes
A short twelve-question self-assessment. What it does not measure: tool knowledge, prompting skill, technical competence. What it looks at: your habits in diagnosing and responding to AI-related behavior in your team.
It is not a psychometric test or an organizational maturity measure; it is a short assessment for awareness. In the program this check is never used alone: it is read together with team interviews, workflow data and leaders’ field tests. Your answers never leave your browser, and no email is required.
The scale: what you do, not what you think
0 / 12The tipping point is between 3 and 4: can you still do it under pressure? Most leadership scales ask about intention; this one asks whether the behavior holds under stress. Some items are deliberately reverse-worded — “5” is not always the right answer.
One loop, three points of intervention
A leader sees an AI-related behavior. Reads it. Makes a move that fits the reading. That move produces a new behavior on the other side — and that behavior seems to confirm the leader’s first reading.
The critical point: this loop runs in the leader and the employee at the same time, and each feeds the other. The employee’s silence grows the leader’s impatience, and the leader’s impatience grows the silence. Nobody acts in bad faith — the system repeats itself.
Not a snapshot, a slope
Two leaders with the same score can be in very different places ninety days later. The real test of the program is not the score on the last day — whether it is still rising three months later.
A plateau is not a failure report; it is a diagnosis. There is almost always a human reason: no reinforcement, no support from the manager, no habit formed. Learning transfer research shows that one of the decisive conditions for persistence is the participant’s own manager’s support.
That is why the program does not end in the room: a weekly check-in card is carried for ninety days, and the participant’s manager receives a separate sponsor brief.
What this is not
Saying clearly what you don’t sell is not a statement of weakness. A buyer who doesn’t know the scope from the start ends up disappointed halfway.
We will work with three organizations in the founding pilot program
We start with a ninety-minute executive briefing: what your adoption dashboard isn’t showing you. Then a 5–6 week pilot: in one or two workflows we choose together, in-depth interviews with your team, a half-day lab with leaders, a two-week field test and a one-page decision report for each workflow. A persistence check ninety days later.
Founding pilot terms: paid, but below the standard rate. In return we ask for two things — structured feedback and, if you agree, a named reference.
Who it’s for: C-level · transformation office · HR leaders · leaders who manage teams