Teams & Organizations
Source, validation and authority are three different things
Short answer
Source tells you where information came from. Validation tells you it was checked. Authority tells you it should serve as the reference. Mixing the three makes organizational memory dangerously ambiguous.

Information can come from the right person and still be wrong.
Information can be true without being a rule.
An official rule can later become obsolete.
In organizational memory, those differences matter.
Source answers: “Where did this come from?”
Provenance helps you know:
- who added the information;
- in which context;
- when;
- whether it came from an import;
- whether it was summarized or directly provided.
That is essential for understanding memory.
But good provenance does not prove truth.
Validation answers: “Did we check this?”
Information can be marked as verified after human confirmation or after checking the system that carries the source of truth.
That changes its confidence level.
It does not necessarily turn it into a rule.
For example:
“The supplier states a 12-day lead time.”
That may be verified.
It does not automatically mean:
“We must order 12 days in advance.”
The second sentence is a decision rule.
Authority answers: “What stands as the reference?”
Some information should serve as the organizational reference:
- a policy;
- a business rule;
- an official decision;
- an approved trade-off.
Its status is different from a shared note or an observation.
Most importantly, authority should not be silently granted by AI.
A concrete example
In a team, a note says: “Discounts capped at 15%.” Who wrote it? A salesperson, in a meeting: that is the source. Did anyone check it with management? Yes, the following week: that is validation. Is it company policy? Only if management adopted it as a rule: that is authority.
Three questions, three possible answers. An AI that does not tell them apart will present the note as a rule.
Why this improves answers
Imagine an AI retrieving three items:
- an idea that was proposed;
- a decision approved later;
- a recent observation suggesting the decision may need review.
If all three have the same status, the AI has to guess.
If their statuses are explicit, the answer can become much more precise:
“The current rule is X. It was decided because of Y. A recent observation may justify reviewing it.”
That hierarchy is what turns a collection of notes into usable context.
Authority must be able to change
Being authoritative does not mean being eternal.
An official decision can be replaced.
Memory then needs to preserve both:
- what stands now;
- the history of what stood before.
The reason for the change often lives in that gap.
The difference between a memory, a fact, a decision and an instruction is detailed in the difference between a memory, a fact, a decision, a hypothesis and an instruction. For replaced decisions, see when a decision changes, what happens to the old one?, and for simultaneous edits, what happens when two people edit the same memory?. The Enterprise and Security pages set out the framework.
Limits of this model
Three statuses do not cover every case, and a real organization has its gray areas: a verbal rule, a practice never written down, a disagreement between two managers. A status model helps ask the right questions, it does not replace an organization’s governance.


