An AI interviewer takes over the first-round interview: the same 10 to 15 minute conversation your team has with every applicant before deciding who goes further. Set up well, it gives every candidate the same questions, the same fair scoring, and an interview slot at any hour. This guide shows how to build that process step by step.

What the first round is for

Be clear about the job of the first round before you automate it. It is not the final interview. It answers three questions:

  1. Does the candidate meet the basics of the role?
  2. Can they explain their experience clearly?
  3. Is there enough evidence to spend a hiring manager's time on them?

Anything beyond that belongs in later rounds with people.

Step 1: Choose three to five competencies

Pick the competencies that matter most for the role, and keep the list short. For a customer support role, that might be communication, problem-solving, and customer empathy. For a sales role: discovery, objection handling, and resilience.

Each competency gets its own score on the scorecard, so a short list keeps the result easy to read.

Step 2: Write questions that ask for real examples

Good first-round questions ask candidates to describe something they did, not what they would do in theory:

  • "Tell me about a time a customer was angry. What did you do?"
  • "Walk me through the last target you missed. What happened next?"

An AI interviewer adds follow-ups based on the answer. If a candidate says "we fixed it", the follow-up asks what they did personally. This is the main difference from a fixed script, and it makes rehearsed answers much less useful.

Step 3: Define the rubric before the first interview

For each competency, write what a strong, average, and weak answer looks like. This keeps scoring consistent from the first candidate to the last, and it gives your team a shared language when they review results.

Step 4: Decide how candidates move forward

Choose one of two modes:

  • Score threshold. Candidates above a set score move to the next stage automatically.
  • Manual review. Everyone lands on the scorecard, and a recruiter picks who goes on.

Start with manual review for the first few roles. Once you trust the scores, a threshold saves more time.

Step 5: Keep the candidate experience simple

Most drop-off happens before the interview starts. Remove friction:

  • One link, sent by email or WhatsApp
  • No app to download; it works in a phone or computer browser
  • Voice or text, so candidates can choose what suits them
  • A short, honest intro that says what the interview is for and how long it takes

Step 6: Review evidence, not just scores

A score alone is not enough to decide. With Inhire, each scorecard shows a score per competency, the full transcript, and audio of key answers, and integrity checks run throughout the interview (for example, tab-focus changes and response timing). Reviewers can read or listen to exactly what earned each score.

Common mistakes

  • Too many questions. A long first round loses candidates. Keep it short.
  • Vague competencies. "Good attitude" cannot be scored. Describe the behavior you want to see.
  • No human check. Review a sample of scores every week, especially at the start.

Where Inhire fits

Inhire's AI interviewer runs the first round by voice or text in English, Bahasa Indonesia, and Bahasa Melayu, asks follow-up questions, and delivers a scorecard with transcript and audio. Your team meets only the candidates worth meeting. Start free and run it on one open role.