01 / Section
Start with the recorded evidence
A captured batch is the starting point, not the conclusion. It may contain documents, messages, pages, files, and other context associated with an attempted business outcome. Describe only what is present and how the items are connected.
Citations make that description traceable. They show where a fact came from, so another reviewer can inspect the same source. A citation proves provenance, not certainty: the source may be incomplete, outdated, ambiguous, or open to more than one reading.
- Recorded evidence: what was captured, where it came from, and what relationship is directly visible.
- Do not turn the presence of a source into a claim that a person read, approved, completed, or relied on it.
02 / Section
Separate interpretation from observation
Interpretation begins when you explain what the evidence may indicate. For example, several files and messages may suggest that a request moved through intake, review, and handoff. That is a useful working interpretation, but it is not the same as observing each step occur.
Use language that keeps the boundary clear: “the evidence shows,” “this may indicate,” and “one plausible sequence is.” Record competing explanations when they matter. A missing artifact can mean a step did not happen, happened elsewhere, or was simply not captured.
- Observation: a review document and a follow-up message appear in the same captured batch.
- Interpretation: the batch may represent a review-to-handoff attempt.
- Not established: that the review was complete, that the handoff was accepted, or that the sequence is the standard process.
03 / Section
Compare attempts, not isolated clues
A single batch rarely explains how work actually happens. Reconstruct distinct attempts at the same or similar business outcome, then compare their evidence. Repeated handoffs, approvals, rework, or missing context can reveal a recurring workflow pattern without claiming that every instance follows it.
Keep each attempt identifiable. Combining separate attempts too early can create a tidy process story that no individual record supports. Differences are often as useful as similarities: they may show alternate routes, exceptions, or points where the available context becomes thin.
- Group attempts by the outcome they appear to address, not only by application or file type.
- Treat recurring patterns as hypotheses to validate with process owners and additional records.
- Do not infer verified working time, effort, productivity, or savings from the shape of a captured sequence.
04 / Section
Use app and site sequences as discovery signals
An app or site sequence can help locate where context changed or where a handoff might have occurred. It is a discovery signal for finding relevant evidence, not a complete account of business work. A sequence does not establish intent, attention, completion, or quality.
When planning a process change, state the assumptions separately. For example, you might assume that a repeated approval step is required, that a named owner can confirm the interpretation, or that a missing system record can be retrieved. Those assumptions should guide validation rather than be presented as findings.
- Evidence: the captured sequence includes a set of applications or sites.
- Interpretation: the sequence may point to a transition between research, preparation, and submission.
- Planning assumption: a process owner can confirm whether that transition is required and what outcome counts as complete.
05 / Section
Make the boundary useful
The goal is not to eliminate interpretation. Operations leaders need working explanations to decide what to investigate, standardize, or redesign. The goal is to make the explanation auditable: show the evidence, label the interpretation, and list the assumptions that still need confirmation.
A practical review ends with questions that can be answered by people or additional records. Which step is explicitly required? Which artifact proves completion? Where do alternate routes begin? What evidence would change the current interpretation? This keeps captured context useful without turning it into a claim that exceeds what the record supports.