Help articles re-checked when the product changes

The decision tree

5 typed decisions, each with an action for every answer

At fixed moments in a run, the engine puts one narrow question to a decision model. The decision model never writes text: it answers yes or no with a probability, picks from listed options, or gives a score, about a small slice of the material. The engine then does exactly what this tree says, which is what makes the run auditable. The thresholds are the template's design values, not measured results.

1. 1

   Planner, while planning

   Scope checkA choice among options

   Before work starts on a unit

   Is this release note specific enough to judge help articles against?

   Sees only: One release note entry

   Why: A vague note cannot prove an article wrong, so it becomes a question for product.

   • Names the changed behaviour precisely→thenAccept
   • Vague, such as general improvements→thenAsk a person
   • Internal change with no customer effect→thenSkip this unit
2. 2

   Scope checkYes or no, with a probability

   Before work starts on a unit

   Does this article or macro describe the screen, setting or step that the release note says has changed?

   Sees only: The release note entry and the article's headings and step list

   Why: Most articles do not depend on most changes, and those are never re-read.

   00.25 and 0.551

   • Yes: 0.55 or higher→thenAccept
   • Unsure: 0.25 up to 0.55→thenAccept
   • No: below 0.25→thenSkip this unit
3. 3

   After workers, the judge checks

   Evidence checkYes or no, with a probability

   After a worker answers

   Does the quoted article sentence tell the customer to do something that the quoted release note says no longer works that way?

   Sees only: The quoted article sentence and the quoted release note

   Why: A knowledge manager flooded with false flags stops reading them.

   00.50 and 0.851

   • Yes: 0.85 or higher→thenAccept
   • Unsure: 0.50 up to 0.85→thenMark unresolved
   • No: below 0.50→thenReject and retry
4. 4

   Run control, between rounds

   Another round?Yes or no, with a probability

   Between rounds

   Are there articles linked to the changed behaviour that have not yet been compared with this release note?

   Sees only: The list of dependent articles with their check status for this release

   Why: Ends the release's re-check once every dependant has been looked at, and no sooner.

   00.20 and 0.501

   • Yes: 0.50 or higher→thenContinue
   • Unsure: 0.20 up to 0.50→thenContinue
   • No: below 0.20→thenStop
5. 5

   Accountable person, before anything is settled

   Person decidesYes or no, with a probability

   Before anything is reported as settled

   Could the release note and the article both be read two ways, so that product must confirm what shipped before a correction is drafted?

   Sees only: One verified flag with both quotes

   Why: Nothing is published automatically, and only product knows what actually shipped.

   Accountable: The knowledge manager approves every published change; product confirms what actually shipped.

   00.10 and 0.351

   • Yes: 0.35 or higher→thenAsk a person
   • Unsure: 0.10 up to 0.35→thenAsk a person
   • No: below 0.10→thenAccept