01 / Section
Repetition is a starting point
Seeing the same kind of task more than once can suggest a recurring workflow. It does not show, by itself, that the work follows a consistent path. Two attempts may share a goal but differ in the inputs, decisions, exceptions, or result.
Before calling something a pattern, define the business outcome you are examining. Then compare separate attempts toward that outcome. The question is not simply whether the same apps appear; it is whether the evidence supports a recurring sequence of work.
02 / Section
Compare attempts, not isolated steps
Review each captured batch as an evidence-linked attempt. Look for steps that recur, where the sequence changes, and what circumstances may explain the difference. Record exceptions rather than smoothing them into a single idealized process.
A useful comparison distinguishes what was recorded from what you infer. For example: “These attempts both include a review before submission” is an observation if the captured evidence shows it. “Review is always required” is a broader claim that may need more evidence.
03 / Section
Treat citations as provenance
A citation lets a reviewer trace a claim to its source in a captured batch. It establishes provenance, not certainty: the source may show only part of an attempt, and an interpretation can still be wrong.
Keep three things visibly separate in your notes: recorded evidence, your interpretation of that evidence, and planning assumptions about what might improve the work. This makes it easier to challenge a conclusion or revise a plan without rewriting what was actually captured.
04 / Section
Use sequences as discovery signals
App and site sequences can help you find attempts worth examining. Treat them as discovery signals, not as proof that a complete workflow occurred or that a sequence caused a particular outcome. Confirm any workflow claim against the relevant attempt evidence.
If the comparison suggests a recurring process, describe its scope and the exceptions you observed. Use that description to decide what to investigate or document next—not to claim more than the evidence supports.
