Where judgment lives

What it is

The AI Survey is a read of where judgment concentrates: which decisions follow a pattern that repeats, and which exist only because one person has made them a hundred times and never written down why. We interview the people doing the work and read the processes underneath it, then hand back a named list of what an agent could carry safely today, what needs more evidence, and what should stay with a person. Nothing here is a score. It is a read a Blueprint can be built from.

What you receive

• 01

  Where the judgment sits

  We interview the people actually doing the work, not only the people managing it, and ask for the last real instance rather than the general policy. Most businesses can describe a process on a whiteboard. Far fewer can say, honestly, which parts of it only happen correctly because one specific person is in the room.

• 02

  What is already written down

  We read the documents, the tickets, and the tools the work actually runs through, and compare that against what the interviews describe. The gap between the two, what people say they do and what the system shows they do, is itself a finding, and it is usually the first thing nobody in the room expected.

• 03

  What an agent could carry today

  We name the processes where the judgment is repeatable enough, and the evidence clear enough, for an agent to take a real first pass under supervision. This is not a wish list. Each one is named against what it would need to be safe: a guard, a reviewer, or a narrower scope than the person currently runs.

• 04

  What should stay with a person

  Just as important as what moves is what does not. Some judgment resists this entirely, not because the technology is not ready, but because the decision genuinely needs a human weighing consequences a pattern cannot see. We name those plainly, so nobody spends a Blueprint's budget chasing the wrong candidate.

We present the findings in person and propose in writing only where the evidence supports a Blueprint. If your real constraint turns out to sit somewhere else, we say so, and the findings are yours either way.

What it needs from you

• Interviews with the people whose judgment is actually in question, not only their managers
• Read access to the documents, tickets, and tools the work runs through today
• A single point of contact who can get us the right people quickly
• Time on the calendar for the work to run, agreed before it starts

How it ends

The AI Survey ends with an in-person presentation and a written proposal, not a sales pitch disguised as one. You leave with a named list of what could move, what should not, and why. If an Agentic Org Chart Blueprint is warranted, the proposal follows the evidence rather than a standard package.