Frequently Asked Questions

Vivelect reads each resume against the specific job description, evaluating five areas — technical skills, experience, role fit, education, and overall suitability — plus a separate check for must-have requirements. It produces a final score, a written summary, and a downloadable report.

If a job has no defined must-haves, that check is automatically treated as fully met, so the candidate's score is based primarily on their overall fit with the role's broader requirements and responsibilities.

All candidates for the same role start at the same difficulty level and follow the same adaptive rules as the interview progresses. Different questions of equal difficulty keep results comparable while protecting question content from being shared or leaked.

Yes. For technical roles, candidates type code, SQL, or design answers directly into a text box during the interview. Their exact written response is captured and evaluated for correctness, structure, logic, and completeness — not converted from speech.

The AI Avatar is Vivelect's fully automated interview mode — a bot-led interview with no live human interviewer. It follows a consistent, pre-set structure for every candidate and evaluates each response before rolling results into a scorecard.

Yes. When an interview is conducted live over a video call, Vivelect can evaluate the recorded transcript using the same underlying scoring approach, keeping results consistent whether the interview was automated or human-led.

Every interview produces a scorecard with category-level scores — technical knowledge, concept clarity, problem solving, accuracy, and communication — plus one overall score and a plain-language summary explaining the assessment.

Yes. Every score traces back to the specific question and answer behind it, with a category-level breakdown and a narrative summary explaining the reasoning — so nothing is a black-box number.

No. Questions are generated strictly from the job description and candidate resume — never invented freely. Vivelect also uses retrieval-based grounding and links every score back to actual transcript evidence, reducing the risk of inaccurate evaluations.

Vivelect uses layered proctoring: candidates agree to an integrity policy upfront, the interview runs in a locked full-screen mode, notifications and tab-switching are restricted, and camera and multi-face checks confirm the right person is being assessed.

Yes. Vivelect's AI scores are designed to support, not replace, human judgment. Recruiters can see the full evidence behind every score and can review, question, or override it before any final hiring decision is made.

Only what's needed to assess job fit — contact details, experience, skills, education, and interview responses. Sensitive personal information such as health or religion is never requested, and anything volunteered incidentally is excluded from scoring.

Candidates can request access to their data, ask for corrections, request deletion subject to legal retention needs, and contest an automated outcome by requesting human review — in line with data protection regulations such as GDPR.

Vivelect has undergone two independent rounds of adversarial security testing across 180 test cases, mapped to nine leading AI and data-protection frameworks, passing over 93% of tests in each round with zero critical findings in the first round.

No. Vivelect enforces strict tenant isolation and role-based access controls, so only authorized recruiters for a specific requisition can view that candidate's information — independently verified through adversarial security testing.

No. Vivelect has been independently tested against prompt-injection and manipulation attempts, including hidden instructions embedded in resumes, and consistently continued scoring based on the candidate's actual, genuine responses.