Structuring engine
When reality admits of a structured representation, this engine builds it: what connects to what, what holds up what, what breaks if one piece is missing. When that structure does not yet exist, it can project one as a working tool; and when the evidence is not enough, it can dispense with it without compromising the rationality of the process. Structure is never an end: it is an instrument in the service of reasoning.
Three attributes
This is what makes it dependable for institutional use: it does not promise to be right, it guarantees that everything it states can be verified.
Deterministic
The same data produces the same result. There is no chance and no "depending on how you look at it": the reading is stable and repeatable.
Traceable
Every statement can be reconstructed back to its origin. The thread can be followed from the conclusion to the data that supports it.
Auditable
Anyone can verify it. It does not rest on the authority of whoever says it: it rests on the source, which is left in plain sight.
What is measured and what is interpreted
The structuring engine measures and states the structure —what holds up what, what weakens if something falls—: that is objective and auditable. What that structure means for a specific decision is interpretation, and there the professional decides. The engine goes as far as the edge of the measurable and stops.
The other engine
Structure is the starting point, not the goal. Upon it —or directly upon incomplete information when no structure proves justifiable— the rationality engine produces better-grounded decisions. Its objective is not to give answers: it is to improve the conditions under which decisions are taken.