One Case File, Two Honest Documents

Opening Scene

Ask the Advocate which document is the real one — the slim folder or the thick binder — and the question doesn’t quite make sense. Both came from the same case file. Neither was written to deceive; both were written to be checked, one by a jury weighing a verdict, one by a judge weighing a process. If the closing argument and the brief ever contradicted each other on a matter of fact, that would be a serious problem — a sign the case itself has an inconsistency, not a sign that summarizing is inherently dishonest. But the fact that they look completely different, in length, tone, and structure, is not itself evidence of a problem. It’s the whole point.

This is the piece that trips up teams building executive and analyst versions of the same report for the first time: the instinct that if the two documents don’t look alike, something must be wrong. Something is wrong only if they disagree. Looking different is not disagreeing.

In Plain English

Derivation, not duplication is the right way to think about the relationship between an executive summary and an analyst report. The executive version is not a shortened copy of the analyst version, and the analyst version is not an expanded copy of the executive one. Both are separately authored artifacts, derived from the same underlying data and analysis, each shaped by the job its specific reader needs to do. The test for whether two such documents are both honest isn’t “do they look the same” — it’s “do they agree on every fact they both happen to state, even though they state very different sets of facts.”

The Old Way

Teams that haven’t internalized derivation-not-duplication tend to run into one of these problems:

Treating the executive version as ground truth, with the analyst version as an afterthought “backup” for anyone who asks questions — which inverts the actual dependency, since the analyst version should be the more complete account the executive version is drawn from.
Treating the analyst version as ground truth and the executive version as a lossy compression of it — technically closer to correct, but it invites treating any simplification as an error, when good executive writing is a deliberate editorial choice, not a degradation.
Panicking over surface differences that aren’t actually disagreements — a stakeholder notices the executive deck doesn’t mention a caveat the analyst report includes, and treats that as evidence of a cover-up, when it may simply be a caveat correctly judged non-essential for the decision the executive needs to make.

What’s Changing (and Why AI Is the Reason)

AI-assisted generation makes it far easier to produce genuinely separate, well-shaped versions instead of one document awkwardly serving two audiences. A model can draft a properly verdict-first executive summary and a properly evidence-first analyst report from the same source material, removing the old excuse that two versions cost twice the effort.
This makes the discipline of checking factual agreement between the two versions a real, ongoing task, not a one-time effort. When both documents can be regenerated quickly, someone needs to actively confirm they still agree on every number they both state — that check does not happen automatically just because both came from the same source data.
AI has also made it easier to accidentally introduce disagreement without anyone intending to. If an executive summary gets regenerated after new data arrives but the analyst report doesn’t, or vice versa, the two artifacts can drift out of factual agreement silently — a risk this series returns to directly in the production block ahead.

The Metaphor, Fully Extended

Courtroom Element	Executive/Analyst Reporting Concept
The case file, containing every piece of evidence	The underlying dataset and analysis both reports are drawn from
The closing argument, written for the jury’s job	The executive summary, shaped for a decide-and-act reader
The legal brief, written for the judge’s job	The analyst report, shaped for a verify-and-defend reader
Two documents disagreeing on the same fact	A genuine, serious defect — a sign the case itself is inconsistent
Two documents simply looking different in length and structure	Not a defect — the expected, honest result of serving two different jobs

For Beginners: What to Actually Do

When you build an executive summary, treat it as a separately authored document with its own structure, not a trimmed copy-paste of the analyst report’s paragraphs.
Before treating a surface difference between the two versions as a problem, check whether it’s an actual factual disagreement or simply a difference in what each version chose to include.
Keep a single source of truth for every number that appears in both documents, so “looking different” never has a chance to quietly become “saying different things.”
Get comfortable with the idea that a well-written executive summary and a well-written analyst report should not resemble each other structurally — resemblance is not the goal, agreement is.

For Practitioners and Leaders: The Deeper Layer

Establish a factual-agreement check as a required step whenever both versions of a report are finalized — not a stylistic comparison, a check of every shared number and claim.
Push back on stakeholders who treat any simplification in the executive version as suspicious; teach the distinction between an honest editorial choice and an actual discrepancy.
Make clear, in team process, which artifact is authoritative for which purpose — the analyst report for methodology disputes, the executive summary for decision records — so nobody has to guess which one to trust when a question comes up later.
Recognize AI-assisted generation as an opportunity to invest saved time into checking agreement between versions, not just into producing more versions faster.

Quick Recap

The executive summary and the analyst report are both separately derived from the same case, not a copy and its shortened version of each other.
Looking different is expected and healthy; disagreeing on a shared fact is the real defect to watch for.
AI has made producing well-shaped versions of both documents cheaper, but agreement between them now has to be actively checked, not assumed.
Treat factual agreement, not structural resemblance, as the actual measure of whether two versions of a report are both honest.

Where This Fits in the Series

This article closes out the foundational block by settling the question every later technique in this series depends on: two differently shaped documents from one honest case, checked for agreement rather than resemblance. Article 5 begins the core-technique block by getting specific about what belongs in the executive version: one page, one number, one action.