An ai interview evidence slot driven follow-up consistency verification method and system

By performing structured analysis and slot filling on candidate answers during AI interviews, detecting slot gaps and generating follow-up questions, the computational, convergent, and replayable nature of multi-round question-and-answer processes in AI interviews is achieved, solving the problems of unverifiable and uninterpretable evaluation results in existing technologies.

CN122240796APending Publication Date: 2026-06-19HUNAN BANYE TECHNOLOGY CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
HUNAN BANYE TECHNOLOGY CO LTD
Filing Date
2026-05-19
Publication Date
2026-06-19

AI Technical Summary

Technical Problem

In existing AI interview technologies, candidate answers lack verifiable elements, follow-up questions lack direction and operability, cross-round consistency is difficult to guarantee, and scoring criteria are difficult to explain and review, resulting in assessment results that are neither verifiable nor reviewable.

Method used

By performing structured parsing of candidates' answers, extracting the claim set and filling it into the evidence slot set, detecting slot gaps, generating a target slot set for follow-up questions, performing cross-round consistency checks, calculating the evidence quality vector, and generating a structured evaluation report.

Benefits of technology

It achieves computability, convergence, and replayability of the multi-round question-and-answer process, ensuring the verifiability and interpretability of the evaluation results, and solving the problems of uncomputable evidence elements, difficulty in convergence of follow-up questions, and unverifiable consistency across rounds.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122240796A_ABST
    Figure CN122240796A_ABST
Patent Text Reader

Abstract

This invention discloses a method and system for verifying consistency of follow-up questions driven by evidence slots in AI interviews. The method includes extracting a set of claims from the candidate's answer text, filling it into the evidence slot set for the target position to obtain a slot filling result, performing slot gap detection to obtain a slot gap set, generating a target slot set for follow-up questions in round t+1 based on the slot gap set, performing cross-round consistency verification based on the slot filling result of round t to obtain a consistency result, and if a preset termination condition is not met, generating the follow-up question object and corresponding follow-up text for round t+1 to proceed to the follow-up questioning round; otherwise, calculating the evidence quality vector, the job dimension score of the target position, and generating a structured evaluation report output. This invention aims to solve the problems of incalculable evidence elements, difficulty in convergence of follow-up questions, unverifiable cross-round consistency, and insufficient interpretability of scores in existing technologies, ensuring that the multi-round question-and-answer evaluation process of AI interviews is calculable, convergent, and replayable.
Need to check novelty before this filing date? Find Prior Art