The first responsible use decision is where not to use it.
Nothing automated reads the career file you submit to CVPRO. No parser, no scoring engine and no language model interprets your record or decides what it means.
Four judgements stay with a person.
The service is designed around an accountable reviewer because these decisions require context, challenge and a source trail.
What a claim means
A title or number can mean very different things across organisations. The reviewer asks what was owned, what changed and what source supports the answer.
What stays in
A claim is not retained because it sounds strong. It stays because the record can support it and the person is prepared to stand behind it.
What the target role asks
Role alignment is a judgement about the mandate in front of the reader. It is not a keyword match against an advert.
What is released
A person reads every draft before delivery and remains responsible for the wording that reaches you.
If the workflow changes, this page changes first.
A future tool cannot quietly become part of career-record interpretation. Any material change must be described in plain language before it is introduced, including what data it receives, why it is used, who reviews the output and how a person can object.
Current position: no parser, automated scoring engine or language model reads submitted career records. The browser-only self-check reads only the answers selected on that page, sends nothing and stores nothing; it deliberately returns a band rather than a Profile Signal Index.