AI assists. People remain responsible.

• Extracting structured information from CV text.
• Extracting requirements from a job description.
• Classifying, for each requirement, what a CV shows, with a quote from the CV as evidence.

The final grouping of candidates, the wording of the summary, and the suggested questions are produced by fixed rules in code. The AI proposes per-requirement classifications; it never writes the verdict.

• It does not independently verify employment, identity, qualifications, work rights, licences, references, or achievements.
• It does not reject, exclude, or hide any candidate. Every uploaded CV stays visible with its evidence, including files that could not be read.
• It does not produce scores, percentages, or a single ranked order presented as objective.
• It is not designed to infer personality, emotion, honesty, health, ethnicity, religion, disability, sexual orientation, political opinion, or other protected characteristics, and personal matters a CV need not mention are never counted against a candidate.
• It does not make or execute a hiring or rejection decision.

• Shown in CV: the CV contains the quoted words. The claim is the candidate's own and is not verified.
• Related experience: the CV shows something genuinely adjacent, quoted, but not the exact requirement.
• Not clear, ask: the CV does not settle it. This is an open question, never a fault, and PickMe suggests the question to ask.
• Not shown: the CV did not provide support. This does not prove the candidate lacks the qualification.
• Needs your eyes: a file PickMe could not read or that duplicated another. Nothing in it was judged.

PickMe Crew is decision support for a human, and several safeguards are built into it rather than left to policy: no automatic rejection, no invented facts, evidence quotes behind every claim, silence reported as unknown, and unreadable files surfaced for human review. Even so, you remain responsible for how you hire:

• Do not use PickMe as the sole basis for rejecting a candidate. Read the evidence, ask the suggested questions, and apply your own judgment.
• Verify material qualifications and claims using fair, job-related methods before relying on them.
• Do not use protected characteristics in your requirements, and provide any legally required notices about AI-assisted recruitment in your jurisdiction.
• Depending on where you operate, using AI in recruitment may carry additional legal obligations. Larger or regulated organisations should obtain qualified legal review before adopting any AI screening tool, including this one.

Review what PickMe extracted before relying on it, and never convert an uncertain or missing requirement into a claim without truthful support. PickMe's output is preparation material for your own application, not a promise of selection.

AI may misread formatting, dates, negations, multilingual content, abbreviations, scanned or image-based documents, role-specific terminology, or the difference between related and direct experience. It may produce fluent text that is incomplete or wrong. PickMe reduces these risks by quoting the CV, refusing to score unreadable files, and reporting silence as unknown, but it does not eliminate them. Job descriptions may themselves be inaccurate, discriminatory, outdated, or ambiguous.

PickMe sends CV and job content to a paid third-party AI provider to perform reading and matching. The operator may use more than one provider, including as an automatic fallback if the primary provider is unavailable, and may change providers over time. Provider requests may be retained temporarily for abuse monitoring under provider rules. See the Privacy Policy for the currently configured providers, links to their documentation, and what PickMe does and does not store.

Use the contact information on the Trust and Safety page to report fabricated evidence, a harmful or discriminatory output, a privacy concern, or a screening you believe treated a candidate unfairly. Include the report or a screenshot and a concise explanation, but do not send identity documents or other unnecessary sensitive information.