Work out what AI actually changed about your job — and what it did not.
A day that replaces both panic and dismissal with a clear read on which parts of managing got cheaper, which got harder, and which decisions are still irreducibly yours.
Everything your job did last month, sorted into cheapened, unchanged and newly hard. Most managers have never looked at the list.
How to tell an artefact that is good from one that is merely plausible — and what you needed to know in order to tell.
What your team will and will not hand to a model, decided deliberately rather than discovered after an incident.
Who it is for
- Managers who feel behind and have been reading tool tutorials instead of thinking about the job.
- Managers who have decided it is hype — the more expensive of the two failure modes, and the harder to talk to.
- Leadership teams who need one shared read before they set a policy.
- No prior technical depth assumed. This is a management day, not a tooling day.
The shape of it
- Format
- One day, on site or remote
- Shape
- Half taught, half worked
- Group
- 8–20 in a cohort; no floor in-house
- Drawn from
- The AI-Native Core
I hold a CTO seat day to day. That is the point: you are not being taught by someone describing a job they left, and the material is drawn from a job I am still doing.
What actually happens
Half taught, half worked — and the worked half runs on your own material, because a case study lets everyone stay polite.
- The skill you were hired for just got cheap
What became abundant, what became scarce, and why those are different questions. The only part of the day that is about the technology.
- What a manager is actually for
Decision rights, accountability, context transfer. None of it was automated, and all of it got harder to do well.
- Judging what the machine produced
The four failure modes of a generated artefact, why non-determinism breaks your testing instincts, and how to build a habit against confident wrongness. Why the same prompt twice is not the same answer twice, what an evaluation set is and why it replaces the test suite you are used to trusting, and the difference between a model that knows something and a system that looks it up — which is why just give it our documents is a project rather than a prompt.
- Why the demo worked and the product did not
The gap every organisation is currently walking into: a prototype that convinced a room in a week, and a year of work between it and something a customer can rely on. What actually sits in that gap — evaluation, the unglamorous data work, cost and latency at real volume, and who is accountable when it is confidently wrong in front of a customer.
- The decision audit
The lab. You leave with your own role on paper and a line you have written, dated and defended.
What you leave with
- A completed decision audit for your own role.
- A rubric for reviewing AI-generated work in your discipline.
- A draft team policy on what does and does not go to a model.
- A follow-up pack with the templates.
What it costs
A cohort seat is €500–600 per person, excluding VAT. In-house starts at €6,000 and is about €8,000 for a room of twelve — and the policy you draft on the day is then your policy rather than an exercise.
It is not, and I would rather say so than invent an instrument. The health check measures the spread between functions; this day is about what changed inside one job, which that instrument does not read. The two Bridge days carry the before-and-after because they are the ones it fits.
Is this the right thing for your team?
Thirty minutes. You describe what is going wrong; I tell you which of these actually fits, including when the answer is none of them.