AI transparency
What it does, what it cannot do, and what happens to what you share. You should be able to read this in five minutes and know whether you want to start.
The short version
It is a thinking partner, not an assessor
No score, no type, no percentile. It offers an observation and asks whether it fits.
It only knows what you tell it
No web search, no LinkedIn scrape, no assumptions filling the gaps.
It gets things wrong, and you can say so
Correcting it is part of the conversation. Flagging a Read as inaccurate reaches a person.
Your content is never sold or trained on
Encrypted in transit and at rest, exportable as a PDF, deletable in full.
Voice is optional and reviewable
Audio is transcribed and not stored, and the text waits for your edit before it is submitted.
There are safeguards for heavy content
The conversation slows for hard material and pauses outright for crisis, with resources offered.
“You remain the authority on your own experience. The AI reflects; you decide.”
ProHumn is a guided career identity experience powered by AI. In one conversation it helps you surface how you are built to contribute: the pattern that runs through your work, in your own words. What you walk away with is your Read.
It is a thinking partner, not a career counselor. It has no authority over your experience and it is not built to acquire any.
And what it is not
ProHumn uses a large language model with specific instructions written for this one conversation. Four things describe most of what it does.
A guided sequence
The conversation moves through a light structure: what came before your career, the roles you have held, and how you are built to contribute.
Pattern recognition
It notices what persists across your stories, the contribution that shows up regardless of title or context.
Reflects, not prescribes
It offers an observation and asks whether it resonates. Your answer is the deciding one.
Adapts to what you share
When heavier content emerges it slows down and gives that space rather than pressing on with the next question.
Every AI system has limits. These are ours, stated before you start rather than in a footnote after you finish.
Pattern recognition is not perfect
It may name a pattern that does not fit, or miss one that matters. Trust your own sense of what resonates over its confidence.
It only knows what you share
Context you do not mention is not considered. It does not search the web, read your LinkedIn, or fill gaps with assumptions.
Cultural context may be limited
The model was trained largely on English-language text and may not fully understand every cultural context around work and identity.
It sometimes gets things wrong
Correct it. Saying "that does not feel right" is a normal and expected part of the conversation, not a failure of yours.
ProHumn makes observations about how you are built to contribute, drawn from what you tell it. Any AI system can carry bias: in the model underneath, in how the questions are framed, and in how it reads different ways of telling a story. We treat that as a design requirement rather than a disclaimer.
Bias is not a problem a product declares solved. What we can state is the practice: assess, document, and change the conversation when a finding warrants it.
Voice input is optional and off unless you choose it. When you use it, this is the exact path your words take.
The short form is here; the full detail is in the Privacy Policy, which is written to be read rather than survived.
Encrypted
Your reflection content is encrypted in transit and at rest.
Never sold, never trained on
Your reflections are not sold to third parties, not used for advertising, and not used to train AI models.
Exportable
Export your Read as a PDF at any time. Your words, your document, your data.
Deletable
Delete your account and everything behind it whenever you decide to. No retention offer stands in the way.
Talking about a decade of work reaches real things: layoffs, failures, grief, people who treated you badly. These safeguards exist because that is predictable, not exceptional.
Trauma-informed pacing
When heavy content emerges the conversation slows and makes room for it instead of steering back to the questionnaire.
Crisis detection
If you express thoughts of self-harm or crisis, the conversation pauses and crisis resources are offered immediately.
No manipulative patterns
No dark patterns, no false urgency, no guilt-based retention, no streaks. Cancelling takes one control.
Clear boundaries
This is a thinking partner for career identity, not therapy. When a conversation moves beyond that scope, the AI says so and redirects.
If your Read does not feel accurate
Tell us it is wrong. That is a more useful outcome for us than your silence.
Use the chat icon at the bottom right of any page and flag the Read as inaccurate. Every flag is reviewed, and it feeds directly into how the conversation is changed. We would rather hear that we misread you than have you carry an assessment that is not true.
Questions about any of this
If something on this page is vague, that is worth telling us about. A transparency statement nobody can follow is not transparency.
Anything on this page
hello@prohumn.com
Flag an inaccurate Read
Chat icon, bottom right of any page
Full data detail