We work with leaders before commitment hardens into momentum. The role is not to supply certainty where none exists, but to make uncertainty visible, distinguish facts from assumptions, and clarify what must be true for a decision to be defensible.
The hardest part of an AI initiative is rarely the technology.
It is knowing what decision must be made, which assumptions deserve to be challenged, and what evidence would justify action. is a Decision Engineering Advisory for consequential AI decisions—when the choice matters, the evidence is incomplete, and accountability cannot be delegated to a model, vendor, or framework.
COMPASS is the Decision Navigation method used to frame the decision, examine the current reality, compare credible routes, test evidence, prepare governance, and decide what should happen next. It is designed to protect judgment—not replace it.
The quality of a decision rarely exceeds the quality of the question that framed it.
The first responsibility in a consequential AI decision is to improve the question. A well-framed question changes what evidence is relevant, which options are credible, who must be involved, and what accountability actually means.
What decision are we actually making—not merely what system are we building?
What would have to be true for this decision to create the intended value?
Which evidence would change our mind, and which assumptions are we currently protecting?
Who remains accountable if the model performs well but the decision still produces harm?
What are the credible alternatives, including not proceeding or changing the problem?
How will we know that value is still being created after launch—not just that the system is still running?
Preserve human ownership where the consequences remain human.
Technology can extend analysis, reveal patterns, and improve foresight. It cannot absorb accountability. At , our mission is to help people make decisions they can explain, govern, revisit, and stand behind.
Prediction can be delegated. Accountability cannot.
Every consequential AI capability becomes part of someone’s decision. It may narrow attention, change confidence, reorder alternatives, or automate a recommendation. The organization still remains answerable for deciding to use it, for defining its authority, and for accepting the consequences.
That is why focuses on the decision around the capability—not only on the capability itself.
Why ?
“vi” means “we” in Swedish. It signals that consequential decisions are rarely owned by one function, one discipline, or one person. “COMPASS” represents orientation when the route is uncertain. The final “a” gives the name an active form: not a static framework, but the practice of navigating together toward a decision that can withstand scrutiny.