---
description: AI Manager Clone workflow – steps to create, configure, and use manager clones
---

# AI Manager Clone Workflow

## Overview
Create a reusable AI interviewer that mimics a hiring manager's style, generates adaptive interview questions, evaluates candidates, and produces structured reports.

## Steps
1. **Manager Profile Configuration**
   - Define manager clone with `name`, `role`, `department`, `expertise`, and `interviewing_style`.
   - Set tone, difficulty, question types, and interview duration.
2. **Knowledge Source Upload**
   - Upload interview guides, question banks, scoring rubrics, and approved sample transcripts.
   - System extracts manager‑specific interviewing patterns and competencies.
3. **AI Persona Generation**
   - Generate a structured manager persona using LLMs.
   - Configure communication style, questioning strategy, and evaluation criteria.
4. **Interview Blueprint Generation**
   - Generate role‑specific interview stages and questions.
   - Define competencies, mandatory questions, scoring weights, and time limits.
   - Provide UI to preview, edit, regenerate, or remove questions.
5. **Adaptive Follow‑Up Questions**
   - Analyze candidate answers in real‑time.
   - Generate follow‑up questions for incomplete, unclear, or incorrect answers.
   - Enforce follow‑up limits and interview progression rules.
6. **Manager‑Specific Evaluation**
   - Evaluate answers using the manager’s approved scoring rubric.
   - Assess technical correctness, problem‑solving, knowledge depth, and completeness.
   - Return evidence‑based feedback and competency scores.
7. **Interview Integration**
   - Integrate with existing text, voice, and video interview modes.
   - Attach selected manager clone to a job or interview session.
   - Maintain consistent configuration throughout the session.
8. **Scorecards and Reports**
   - Generate structured evaluation scorecards.
   - Summarize candidate strengths, weaknesses, competency gaps, and recommendations.
   - Export PDF reports for recruiter review.
9. **Testing and Version Management**
   - Simulate interviews before publishing.
   - Test question quality and scoring consistency.
   - Support draft, publish, archive, version history, and rollback.
10. **Security and Access Control**
    - Restrict clone creation and editing to authorized users.
    - Enforce company‑level data isolation.
    - Maintain audit logs, consent, and source‑data privacy.
11. **AI and Technical Implementation**
    - Use LLMs for persona extraction, question generation, follow‑ups, and evaluation.
    - Use Retrieval‑Augmented Generation (RAG) to fetch relevant manager knowledge.
    - Integrate with existing FastAPI backend, session lifecycle, and evaluation services.
12. **Quality Assurance**
    - Test API validation, interview workflows, scoring, and error handling.
    - Verify tenant isolation, prompt‑injection protection, and session consistency.
    - Compare AI evaluations against qualified human reviewers.

## Expected Outcome
A configurable, reusable AI interviewer that follows an approved hiring manager's approach, asks adaptive questions, evaluates candidates using a consistent rubric, and generates structured reports for recruiter review.
