A service that sells credibility has to show its working. If anything below is wrong, we would rather hear it than not.
Xpert Lens reads behaviour demonstrated during an assessment. It is not a personality test, and there is no type, profile or colour at the end of it.
What is measured, and what counts as evidence for it, comes from twenty-five years of interviewing candidates professionally — not from a model deciding for itself what mattered. More on that here.
Each area comes back as one of three bands. No numbers out of ten, no percentiles, no scores — no assessment of this kind can support that precision, and a number invites a false comparison between two people who answered entirely different questions.
There is one number in the product and it is a different kind of thing. JD Doctor counts requirements — how many of the things a job description asks for your CV covers. You can read the list it counted and disagree with any line of it. That is arithmetic about two documents, not a verdict on a person, and it is never a probability of being hired.
Nothing that could score an answer is ever sent to your device. No answer key, and no indication of what any option signals — your browser genuinely does not have the means to work out your result.
Our servers resolve every signal, derive every band, and write the credential from their own output. The questions you were served are recorded when the assessment begins and cannot be altered afterwards.
Why this matters more than it sounds. If scoring happens in the browser, then "verified" means the candidate's own device said so — and the single thing being sold is trustworthiness, so one person noticing takes down the entire proposition rather than one feature. We changed this before taking a single payment.
Most people taking this assessment are working in their second or third language. Several of the things a careless assessment would penalise are not evidence of capability at all — they are evidence of where somebody grew up.
Speech timing is measured, and it is used for one purpose only: deciding whether there is enough speech to assess at all. It never sets a band.
An assessment is structured evidence. A conversation about yourself is self-reporting again, however well it is written up. A CV that says "led a turnaround" because the candidate mentioned it in a chat is not validated — it is laundered, and it looks validated because it sits next to things that are.
So the rule, which the product enforces rather than promises:
The same principle runs through the assessment itself. What you say aloud can earn a band; what you type afterwards can correct what you said, but never buys a delivery band you did not demonstrate.
Your assessment, your report, your CV and your credential belong to you. Nothing is shown to an employer or a recruiter unless you turn sharing on, and you can turn it off again — a credential you have withdrawn stops verifying.
Your CV is read to produce your own report and your own documents. It is not added to any searchable database, and it is not offered to recruiters.
You can delete the voice recording from an assessment and keep the transcript.
The full position is in the privacy policy.
The methodology is meant to be argued with. That is the point of publishing it.
Write to us