The Samples That Teach Kretell the Most About You Are Not on LinkedIn
Maya has a Substack with 2,100 subscribers. She has written for Forbes Councils. Her LinkedIn posts, at their best, were among the most engaged content in her professional network.
Then she adopted AI tools. Her output doubled. Her engagement halved. The specific bicultural frame that made her perspective rare — the vantage point of someone who operates across African market infrastructure and European investment strategy — had been averaged out.
She came back to Kretell to reclaim it. She uploaded her LinkedIn posts. The Voice Profile was accurate — better than any other tool she had used.
But not quite back to the Maya her subscribers remembered.
The reason was in the samples.
Why the Professional Mask Produces a Less Accurate Profile
The Voice Profile learns from patterns that repeat consistently. If your public writing reflects your curated voice rather than your actual voice, the Profile learns the curation. That is what it has to work with.
When the output generates, it generates from what it learned. The output sounds like your curated professional voice — competent, polished, well-structured. Which is exactly what every AI tool produces.
The distinction between your curated voice and your actual voice is not large. But it is the distance between content that sounds professional and content that sounds like you specifically.
Maya's readers subscribed to her because they recognised her specific bicultural frame — the way she places African market context against European strategy analysis. That frame is most fully present in the writing she produced when she was thinking rather than performing. In the Substack drafts she wrote before she edited them for readers. In the working documents she shared with peers.
Those are the samples that close the gap between a good Voice Profile and the Mirror Moment.
The Three Sample Types That Improve Profiles Fastest
Substantive peer correspondence. Emails and messages to colleagues whose expertise you respect, written with full precision because they would notice if you glossed. These contain your real argumentative patterns — how you build a case, how you acknowledge uncertainty, how you credit others when you are working through a shared problem.
Unpolished working documents. Notes, internal briefs, project summaries written for function rather than audience. The vocabulary you use when you are not considering how it will be received is more consistently yours than the vocabulary you deploy for public writing.
High-stakes private writing. The pieces you wrote when the outcome mattered more than the impression. Professional references, detailed feedback documents, analyses where being right was more important than being impressive. Your voice is most distinctly present when you are focused on the substance rather than the form.
Add these to your existing Voice Profile samples. The system will identify the patterns that appear across both your public and private writing — those are the patterns most deeply yours.
The Calibration Behind the Voice
Kretell's cultural intelligence calibrates to the communication register of your professional market. Launched across 19 markets — with every market researched natively, not assumed. The calibration covers self-promotion register, formality, directness, humility framing, and how expertise is signalled in the specific professional culture you operate in.
For Maya, whose professional identity spans two cultural contexts, this calibration is not a single-market setting. It is a specific combination of registers — the precision of a European financial services professional and the relationship-oriented framing of someone who built credibility in African markets.
That combination is documented in the cultural intelligence layer. The Voice Profile maps the individual patterns. Together, they produce output that sounds like Maya — not Maya averaged into a Western professional default.
When Your Voice Comes Back
The first Kretell post that stops Maya mid-read — not because something is off, but because it sounds exactly like the person her subscribers followed before she let AI average her out — is the Mirror Moment.
It is not a feature. It is the product doing what it was built to do: ensure that every word you publish sounds unmistakably like you.
That moment arrives earlier when the samples you provide give the system access to your real voice rather than your curated voice.
Your voice is still there. Kretell helps you find it again.
Frequently Asked Questions
What writing samples give Kretell the most accurate Voice Profile?
The most accurate profiles come from writing produced without a public audience in mind. Substantive emails to respected peers, internal memos, working documents, and detailed project notes carry your real voice patterns — the ones you use when you are focused on the substance rather than the impression. Public-facing writing like polished LinkedIn posts reflects your curated voice, which is yours but not your full voice. A mix of both types produces the most complete profile.
Why does my Kretell output not sound exactly like me yet?
The most common cause is samples that are too narrow or too polished. If all your samples come from one format — LinkedIn posts only, for example — the profile has a partial picture. It knows how you write when you are performing for a professional audience. It does not yet know how you write when you are thinking precisely. Add internal documents, substantive peer correspondence, and any writing produced for function rather than impression.
How does Kretell learn from my edits?
Every edit you make to a generated post teaches the profile something. The system analyses the delta between what it generated and what you replaced it with. Deliberate replacements — where you change a word or phrase because it is not yours, and replace it with exactly what you would have written — teach the profile the most. Smoothing edits that adjust rather than replace are less instructive. The more specific the correction, the faster the profile adapts.
Can I use private or confidential writing samples with Kretell?
You can use any writing that reflects your authentic voice patterns. Kretell's Voice Sovereignty commitment means your Voice Profile is never shared, sold, or used beyond generating your own content. If you prefer not to upload sensitive internal documents directly, edited excerpts that preserve your sentence patterns and vocabulary without the underlying confidential content work equally well as training material.
Does the Voice Profile ever become less accurate over time?
The profile does not degrade. It deepens. Every generation and every edit adds to the map. The profile can become temporarily less accurate if a user's voice undergoes a significant shift — a career change, a major context shift, years of disuse. In those cases, adding fresh samples from the new context recalibrates the profile forward without erasing the existing foundation.
How is Kretell's voice profiling different from style settings in other tools?
Style settings describe a type of writing. A Voice Profile maps an individual's writing. The distinction is architectural. Style settings apply a category — "formal," "confident," "conversational" — that any professional could fit into. A Voice Profile captures the specific patterns that make one professional's writing distinct from every other professional's: the specific sentence architecture, the specific vocabulary range, the specific cultural register, the specific attribution patterns that belong to that person alone.
