Decision Authority Research
Board Brief
Named Stop Authority
Decision Authority Research  ·  Liran Bar-Shrim  ·  2026

If something goes wrong today, who by name can stop it before the outcome becomes irreversible?

If this question cannot be answered in one sentence, accountability is already undefined.
Most boards can name who approves. Almost none can name who stops.

Organizations invest heavily in approval authority. Budget sign-off, risk committee review, board oversight. These mechanisms are well-documented and well-understood.

Stop authority is different. When a high-stakes decision must be halted before execution becomes irreversible, the question is not who approved it. The question is who, by name, is required to stop it.

When no single name is committed in advance to stop, the default is continuation. Momentum decides. Not governance.
Autonomous systems execute faster than governance convenes.

AI-driven credit decisions, automated trading, clinical support systems, and agentic workflows can produce irreversible outcomes before any human review process has formally begun.

The governance question is no longer who reviews afterward. It is who can halt the system before execution crosses the point of no return.

When no person owns the stop, the system defaults to continuation. If nobody owns the brake, speed becomes the decision-maker.
Stop authority requires three names. Not one role. Not a committee.

Each function is distinct and non-overlapping. The same individual cannot hold more than one.

01
Stop
Named in advance. Required to halt execution. Protected from retaliation for good-faith invocation.
02
Hold
Maintains suspension while governance responds. Accountable for the duration. Cannot release unilaterally.
03
Release
A separate individual. Empowered to resume only after documented review. Release carries personal accountability.
Stop is the default state. Continuation does not exist until explicitly released by a named individual bound to refusal.
Where personal liability concentrates.

When a decision produces harm and no named stop authority existed in advance, accountability defaults to those at the top of the governance chain. The absence of a named stop is not a neutral design choice. It is an exposure.

When an organization claims it has a stop mechanism but cannot produce a name, it does not have a stop mechanism. It has a stop narrative.
Three questions for the board.
01  ·  For your most consequential automated decision today, who by name holds explicit stop authority?
02  ·  If that person invokes a stop, who by name holds it until governance responds?
03  ·  Who by name can release the stop, and what documentation is required before they may do so?
If you cannot name these three roles today, you are already operating without a stop.
Most organizations believe they have a stop.
Very few can name it.
If you cannot, the decision has already been made.
Continuation is your default.
For organizations operating irreversible decision systems,
this question is not theoretical.

Private review available.

Request a Private Stop Review