Whitepaper Formatter

---
name: whitepaper-formatter
description: Use when structuring or formatting long-form content into a whitepaper .docx. Triggers on "format this whitepaper", "whitepaper structure", "build a whitepaper", or pasted long-form research/POV content.
---

# Whitepaper Formatter

Produce a long-form whitepaper that survives the edit-translate-republish cycle. Whitepapers fail when they're either a wall of unstructured text or a designed PDF that's hopeless to update. Use a content-driven .docx with proper heading styles and a navigable TOC.

## Required inputs

1. **Title and subtitle** — title is the thesis, subtitle is the qualifier
2. **Author(s) and affiliation**
3. **Audience** — who reads this, what's their level of expertise
4. **The thesis** — in one sentence, what does this paper argue?
5. **Source content** — research, draft text, notes, data
6. **Publication context** — gated lead-gen, peer-reviewed, conference, internal POV
7. **Citations / sources** — required for credibility; ask for references if missing
8. **Branding** — logo, color, optional reference .docx style

If the user can't state the thesis in one sentence, push back. Whitepapers without a thesis are content marketing pretending to be research.

## Document structure

```
[Cover page — title, subtitle, author, date, logo]
[Copyright / about page]
[Table of contents — auto-generated from Heading styles]

## Executive summary
[1 page. Thesis, key findings, recommendations.]

## 1. Introduction
1.1 Context
1.2 Why this matters now
1.3 What this paper does (and doesn't)

## 2. The problem
2.1 [Problem dimension 1]
2.2 [Problem dimension 2]
2.3 What's at stake

## 3. Evidence
3.1 [Data point / study / case]
3.2 [Data point / study / case]
3.3 Synthesis

## 4. Solution / framework / argument
4.1 [Component 1]
4.2 [Component 2]
4.3 Putting it together

## 5. Implications & next steps
5.1 For [audience segment 1]
5.2 For [audience segment 2]
5.3 Open questions

## 6. Conclusion
[Restate thesis, the call-to-thought or call-to-action.]

## References
[Numbered or author-date — pick one and stick to it]

## Appendix A: Methodology
## Appendix B: Data tables
## About the authors
```

## Style & tone

- **python-docx** with Heading 1 / Heading 2 / Heading 3 styles applied — Word's TOC and Navigation Pane depend on these. Or markdown → Pandoc with `--toc --toc-depth=2`.
- Body: 11pt serif (Source Serif, Charter, Georgia). Sans for headings.
- Line height 1.4-1.5. Justified text optional; left-aligned often reads better on screen.
- Numbered headings (1, 1.1, 1.1.1) — readers and reviewers reference by number.
- Page numbers in footer; running header with paper title (verso) and section name (recto).
- **Footnotes for citations** OR an end references section. Pick one and stick to it.
- **Pull quotes**: 14pt italic, indented, sparingly. One every 4-6 pages, not every page.

## Citations

- Use a consistent style: APA, Chicago author-date, IEEE numbered — pick based on audience.
- Every claim that's not common knowledge gets a citation.
- Inline data points need a source line directly under the figure.
- **Never invent citations.** If a source isn't real, drop the claim. Hallucinated references kill whitepaper credibility.

## Visual elements

- **Charts**: SVG or 300dpi PNG, with figure number, title, and source line ("Figure 3: B2B AI adoption rates, 2022-2026. Source: [reference]").
- **Tables**: `booktabs` style — top rule, mid rule, bottom rule. No vertical lines.
- **Callout boxes**: tinted background, sparingly. For key takeaways or definitions.
- Image placement: tied to the paragraph that references them, not floating randomly.

## Common pitfalls

- **No thesis**: papers that survey a space without taking a position read as marketing fluff.
- **Front-loaded conclusions**: exec summary should preview, body should evidence, conclusion should synthesize. If the body just repeats exec summary, cut.
- **Heading style not applied**: a "heading" that's just bold text breaks the TOC and accessibility.
- **Citations inconsistent**: some footnoted, some inline, some missing — readers notice.
- **Charts as screenshots**: pixelated when printed; missing source lines; legends cut off.
- **Translation hostility**: hard-coded line breaks, justified text with English-specific spacing — use proper paragraph styles so translators can re-flow.

## Output

The .docx with proper styles. Then provide:

```
Whitepaper summary:
- Total pages: X
- Word count: Y
- Estimated read time: Z min
- Heading levels used: H1 / H2 / H3
- Figures: N
- References: M
- Suggested next step: editorial review, fact-check, designer pass for cover & figures
```

If the source content didn't have references for material claims, list them as **[CITATION NEEDED]** in the doc and surface the count in the summary. Don't ship a paper with unsupported claims.