Glossary — AI interview terminology
This glossary is a plain-language reference for terms that come up across modern AI interviewing. Each definition is kept short and concrete; the mechanics, practical hiring implications, and related concepts live on dedicated entry pages. The set is meant for both HR teams and candidate-side readers, with cross-links to academic sources and the relevant areas of the Intrvio platform itself.
Adverse Impact
Adverse impact occurs when a selection method screens out a protected group at below four-fifths the pass rate of the highest-passing group.
Read entryAgentic Interview
An AI-conducted interview where the system adapts its next question based on prior answers — the conversation has agency, not a fixed script.
Read entryAnswer Authenticity
Answer authenticity refers to signals that a candidate's interview responses reflect genuine personal knowledge rather than AI-generated or coached content.
Read entryAsynchronous Interview
An asynchronous interview lets candidates complete their responses without a live interviewer present, typically within a 24-to-72-hour window.
Read entryBehavioral Rating Scales (BARS)
BARS combine numeric ratings with anchored behavioral descriptions per level, so two interviewers score the same answer the same way.
Read entryCandidate Scoring Rubric
A candidate scoring rubric defines expected behaviors at each score level (1-5) per interview question, enabling interviewers to rate answers consistently.
Read entryCompetency-Based Interview
A competency-based interview maps each question to a specific role competency and scores behavioral evidence against a defined rubric.
Read entryInterview Proctoring
Interview proctoring monitors environmental conditions during a remote interview to surface integrity signals for human review.
Read entryStructured Interview
A hiring interview where every candidate answers the same predefined questions and is scored on the same rubric.
Read entryVoice AI Interviewer
Software that conducts interviews via real-time spoken conversation, using STT + LLM + TTS in a low-latency loop.
Read entry
When the glossary is not enough, try a real interview.
Run a structured, multilingual, agentic interview with GAIA in just a few minutes.