01 / Section
Treat evidence as a starting point
A review should begin with evidence linked to a captured batch and a business outcome. That makes it possible to inspect what supports a proposed workflow instead of relying on a summary alone.
A citation establishes where a piece of evidence came from; it does not prove that the evidence is complete or that an interpretation is correct. Treat each recorded attempt as one example to examine, not a verdict about how work always happens.
02 / Section
Compare attempts, not people
Compare distinct attempts at the same kind of business outcome. Look for steps that recur, steps that vary, and points where the evidence is unclear. Those patterns can help a team identify a workflow worth reviewing.
A difference between attempts is a reason to ask a question, not a basis for judging an individual. The goal is to understand the work and its conditions, not to assess employees.
03 / Section
Keep discovery signals separate
App and site sequences can suggest where a workflow might begin or which tools may be involved. Keep these sequences separate from evidence about a specific attempt and its outcome. A sequence can point to something worth investigating; on its own, it does not establish what task was completed or why.
This distinction helps reviewers avoid turning a possible lead into a claim. If a sequence matters to the proposal, state what it suggests and what further evidence or context is needed.
04 / Section
Label evidence, interpretation, and assumptions
Write down three things separately: what the captured evidence records, what the reviewer thinks it means, and what the team is assuming for planning. For example, a recorded sequence may show a handoff between tools; an interpretation may be that the handoff is part of a recurring workflow; an assumption may be that the same handoff applies to other cases.
Make each layer easy to challenge. If an assumption is necessary to estimate scope or design a proposed change, name it explicitly and identify who can check it. Do not present an assumption as an observed fact.
- Recorded evidence: what the batch shows, with its source.
- Interpretation: the pattern the reviewer sees across attempts.
- Planning assumption: what still needs confirmation before a decision.
05 / Section
Put a human decision before automation
Use the review to frame a decision, not to make one automatically. Ask whether the recurring pattern is understood, which exceptions matter, what evidence is missing, and who owns the process. If the answers are unclear, the next step is further review—not approval by default.
A responsible proposal states its scope, open questions, and conditions for proceeding. The people accountable for the business process can then decide whether to test or pursue it, and what safeguards or additional checks are needed. Visible assumptions make that decision more informed; they do not remove the need for human judgment.
