Unable to determine - insufficient data
232
Samples Analyzed
53,247
Total Words
4
Languages
insufficient
Corpus Quality
Writing style compared to average AI output
Intrinsic stylometry metrics
Vocabulary Richness
TTR 0.21
Writing Rhythm
Moderate
Sentence Range
words
Signature Punctuation
Every sample opens not with the content itself but with an explanation of the content — labeling options, describing intent, and previewing what follows before delivering it.
“"Here are a few options, ranging from short and sweet to a little more detailed" and "Voici une proposition de structure... J'ai conçu ce modèle avec des espaces réservés"”
Numbered lists, bracketed placeholders, and closing tip sections appear in nearly every sample, framing outputs as reusable templates rather than finished communication.
“"Conseils pour remplir les champs", "Consejos clave utilizados en este correo", and "ポイント：" sections appear across FR, ES, and JA samples alike”
Rather than committing to a single voice or approach, the writing consistently presents 2–3 labeled variants ranked by tone or relationship proximity.
“"Option 1: Short and standard (Best for someone you know well)" / "Option 3: Si vous avez une relation très proche/collaborative"”
Substantive content is systematically withheld behind fill-in tokens, meaning no sample reflects a completed, authentic personal voice — only a framework for one.
“"[Nom du projet]", "[X]%", "[機能名]", "[Tu Cargo]" appear hundreds of times across the corpus”
The dominant register across all languages and formats is that of a writing coach explaining strategy to a reader, not an individual communicating their own perspective.
“"Soyez honnête : Si les chiffres sont mauvais, ne les cachez pas" and "Don't let the algorithm steal your voice. Use the tool, but keep the soul."”
Clear, accessible language
Measured, professional tone
Conversational, approachable style
Gray line = average AI output. Purple bar = their writing. The bigger the difference, the more distinctive the voice.
All three formal email samples follow the identical structure of multi-option presentation with closing tip sections — no individual voice distinguishes the French from the Spanish from the Japanese output.
Report samples are purely structural templates with dense placeholder tokens; they demonstrate competence with QBR frameworks but contain zero authentic authorial data to analyze.
These are the only samples approaching a personal voice — the AI-generated anecdotes ("I hated it", "trois phrases maladroites, mais sincères") mimic a reflective LinkedIn register, but the stories are fabricated and formulaic across all three languages.
Casual email samples are the shortest and most stripped of personality — each delivers 2–3 subject line variants with minimal body text, offering the least material for voice extraction.
Discover what makes your writing uniquely yours — free Writing DNA Snapshot included.
20% off with code VDN-01A292
Powered by MyWritingTwin.com