Screen every applicant. Reject no one by machine.

A bare score is a guess you cannot check.

Most screening tools rank candidates and stop there. The recruiter cannot see why one person scored 74 and another 61, so the number is either trusted blindly or ignored.

Pyro shows its working. In the example above, the score is high on skills and seniority, and the one open question, which decisions were his, goes to ARIA's first round instead of becoming a rejection.

Knockout questions, keywords, meaning and criteria.

• Knockout questions reject anyone who answers one question the wrong way. Fast, but blunt: one answer decides.
• Keyword filters keep CVs that contain the right words, and miss people who describe the same work differently.
• Semantic matching compares the meaning of a CV with the role, so different wording still matches.
• Criteria-based screening scores each application against criteria written for the role, and shows its reasoning. Pyro combines meaning with criteria, and never rejects anyone on its own.

What a good screening result should show you.

• A score you can compare across every applicant for the role.
• The skills that matched, and the required skills that are missing.
• A view of seniority, domain experience and ownership, in words.
• What is still open, as questions for the interview.
• Nothing that decides for you: the next step is a person’s call.

Criteria that screen well.

Screening is only as good as the criteria behind it. Pyro agrees them with you before the posting goes live, and these habits make them work harder.

• Separate what the role needs from what would be nice. Too many “requirements” put good people off and blur the screen.
• Describe outcomes, not just tools: what the person must have done, at what scale.
• Keep criteria the same for every applicant to the role, so the scores can be compared.
• Leave out anything that is not about the job, such as age or background.
• Write the criteria once and let them carry through: posting, screening, interview and decision.

Screening that is fair to the person applying.

Every applicant to a role is measured against the same written criteria, and nobody is turned away by a machine. A gap in a CV becomes a question in the interview, so a candidate who did the work but described it briefly still gets the chance to explain.

Candidates also see where their application stands as it changes, so nobody is left waiting without word.

Where screening usually goes wrong.

• Rejecting on a single missing word, when the experience is there in other words.
• Ranking without reasons, so nobody can explain the shortlist.
• Moving the bar from candidate to candidate.
• Treating a gap as a verdict instead of a question for the interview.
• Letting software make the final call.

How a screened application reaches you.

• The application arrives through your careers page or the Pyro jobs board.
• Pyro reads the CV and scores it against the role’s criteria.
• You see the score with its reasoning: what matched, and what is missing.
• If the score clears the bar you set, Pyro invites the candidate to an ARIA first round.
• You decide every next step, and your name goes on the record.

Screening that a person can stand behind.

Screening is where hiring law and AI meet most often. Hiring is a high-risk use under the EU AI Act, and anti-discrimination law applies to decisions made with software. Pyro is designed so a person can explain every outcome.

• Every candidate for a role is measured against the same written criteria.
• Pyro never rejects a candidate or moves anyone between stages on its own.
• Diversity data is kept apart: never shown to decision-makers and never sent to the AI.
• Postings are graded for biased wording before anyone applies.

The rules that apply to automated screening.

Several laws now cover software used in hiring decisions. The EU AI Act treats AI used in recruitment as high-risk. In the United States, New York City requires bias audits and notice for automated employment decision tools, Illinois regulates the use of AI to analyse video interviews, and Colorado has passed its own act on high-risk AI systems.

The common thread is explanation and human oversight: being able to say why a candidate was ranked as they were, and keeping a person responsible for the decision. That is how Pyro is designed. Your own legal obligations remain yours, and your legal adviser is the right person to confirm what applies to you.

Screening for agencies and for startups.

Agencies screen for many clients at once, often from a database of past applicants. Pyro scores those CVs against each client’s criteria, so the same person can be weighed for several roles.

Startups usually screen without a recruiter. Pyro gives founders a ranked list with its reasoning, and invites the strongest candidates to a first-round interview, so the founder’s time goes on the final conversations.