Name Standard℠ Advisory
The AI did the work. A human signed it. When someone asks who owned that outcome — will the answer hold up?
The decision arrived already made.
An AI agent monitors inventory, triggers the reorder, renegotiates the vendor terms, drafts the contract, and confirms the budget. Then it hands a human a decision that is already substantially made — and the human signs.
That signature now carries the accountability. But did the person behind it have the time, the information, the authority, and the real option to say no? Or was the sign-off ceremony?
The frameworks built for a human world — RACI charts, approval matrices, delegation of authority — break down the moment AI agents enter the workflow alongside people. A human name on AI-assisted work is not, by itself, evidence of human accountability. The Name Standard℠ tests whether it could be.
Time allocation
Review capacity and tools
Information access
Documentation infrastructure
Authority and escalation
Formal right of refusal
The questions leaders can't yet answer
When AI does the work and a human approves it, who owns the outcome?
Not in the org chart — in fact. If the approval is challenged, can the organization show the named person genuinely reviewed, understood, and chose? We test each attribution against that standard before someone else does.
Could the named reviewer actually have said no?
An approval only means something if refusal was a real option — with time to exercise it and no penalty for using it. Where sign-off is structurally unrefusable, accountability has already diffused. We find those points.
Would the review survive reconstruction?
Board records, professional advice, and regulatory submissions are discoverable if they exist and are relevant — including AI-generated drafts and parallel outputs that never entered the official record. We test whether the human review behind the name can be reconstructed and defended later.
Where has responsibility quietly melted together?
As AI agents span procurement, finance, legal, and operations in a single workflow, the functions that used to own each step blur into one another. Ownership looks clear on paper and diffuse in practice. We map where attribution has become ceremonial.
Six tests behind every name
Attribution review
Locate where AI-assisted work already carries a human or institutional name — approvals, advice, client communications, regulatory representations, board materials.
Capacity test
Did the reviewer have the time and tools for meaningful review — or was the volume of AI output already beyond what any human could genuinely scrutinize?
Information test
Did the named person see the evidence behind the recommendation — or only the conclusion the workflow handed them?
Documentation test
Can the review and the decision be reconstructed later — which steps were AI, which were human, and on what basis the human decided?
Authority test
Could the named person approve, constrain, escalate, or stop the work — including stopping the AI agent itself?
Refusal test
Was declining to sign a real, formal, exercisable option? Human judgment at the decision point is not friction — it is the control layer. This tests whether the layer exists.
A clear advisory process
Identify
Map the AI-assisted decisions, outputs, and approvals that carry a human name — starting where exposure is highest.
Test
Evaluate each attribution against the six Name Standard℠ conditions: time, capacity, information, documentation, authority, refusal.
Locate
Pinpoint where the conditions break down — where a signature is carrying accountability the signer could not actually exercise.
Brief
Deliver a clear view of where attribution is defensible, where it is ceremonial, and what leadership must resolve first.
Common questions
Is this a software audit?
No. It is a governance review of the humans whose names carry AI-assisted work: their authority, evidence, review capacity, and refusal rights. The technology is context; the accountability is the subject.
Is this an argument against using AI?
No. It is an argument that human judgment at the decision point is the control layer that makes AI-driven work governable — and that the layer has to be real, not ceremonial, to protect the organization and the people signing.
What kinds of work can be reviewed?
Executive approvals, professional advice, client communications, regulatory submissions, board materials, financial reports, and any other AI-assisted work where a person or the institution stands behind the output.
Request a briefing
Talk to us about name standard advisory for your organization — no obligation, and nothing shared outside this inquiry.
Thank you!
Your request has been received. We'll follow up to scope the briefing.