Purpose

Use this article to understand the candidate-facing experience for AI interviews.

Invites and interview links

How candidates receive invites

After you send an AI interview invite, Eightfold:

• Sends the candidate an email with a link to start the AI interview.
• May also send an SMS or WhatsApp message, if configured in your environment.

The invite explains that the interview will be conducted by an AI interviewer and directs the candidate to click the link when they are ready.

Link validity

• The interview link remains valid until it reaches the link expiry duration configured by your admin (for example, 7–14 days).
• If the link has expired, the candidate sees an error and cannot join; you must send a new invite.

Identity verification (optional)

If your organization has enabled identity verification for AI interviews:

• After pre‑join checks, the candidate is taken through a CLEAR-based verification flow.
• This typically includes document capture and/or biometric checks, guided by CLEAR’s UI.
• Once verification is complete, they are automatically returned to the AI interview experience.

This step:

• Applies across AI interview types when turned on.
• Is designed to be mobile‑friendly and fast, so candidates can complete it on phones as well as desktops.

If identity verification fails or is incomplete, the candidate may not be allowed to proceed, depending on your configuration.

To know more, read Identity verification in AI Interviews.

Pre‑join systems check

Before the AI interview starts, candidates go through an improved pre‑join systems check flow. This is designed to catch common technical issues and provide simple ways to fix them.

What is checked

The pre‑join screen runs a set of checks, such as:

• Microphone – Is a mic connected and permitted?
• Camera – Is a camera available and permitted (if video is required)?
• Speakers / audio output – Can the candidate hear the AI interviewer?
• Network / VPN connectivity – Can the system reach required services (for example, the speech and audio provider)?

If enabled in the Admin Console, a two-way audio connectivity check is performed with an AI agent.

If any of these fail, the candidate sees clear error messages describing what is wrong (for example, 'Microphone not detected' or 'Network connection blocked').

How candidates fix issues

When an issue is detected, the pre‑join screen:

• Lists the problems that are preventing them from joining the call.
• Provides basic instructions (for example, to check browser permissions or unplug and re‑plug devices).
• Offers a way to retry the checks after the candidate adjusts their settings.

Candidates must resolve blocking issues (like no microphone access) before they can continue into the interview.

When the systems check passes

If all required checks pass:

• The candidate can click Join interview to enter the AI interview session.
• From this point, the AI interviewer takes over: it greets the candidate and explains the format for the chosen interview type.

Experience by interview type

Once health checks (and identity verification, if enabled) are complete, the candidate enters the live AI interview.

Across all interview types, the experience shares some common behaviors:

• The AI interviewer greets the candidate, introduces itself, and explains what'll happen.
• The conversation is conducted in the language of the selected AI agent (for example, English (US), German, Spanish).
• A real‑time transcript may appear on screen, depending on your configuration, so candidates can read what the AI is saying as it speaks.

Profile screening interviews

In a profile screening interview, candidates experience a conversational Q&A that focuses on fit and interest for the role.

How questions are generated

The AI interviewer asks questions based on your scheduling choices (while setting up Ai Interviewer), which the candidate doesn't see directly.

• When you choose Feedback form, job description, and resume, questions are tailored to the role and the individual candidate.
• When you choose Feedback form only, questions are standardized and drawn from the feedback form.

What candidates experience

During the interview, candidates:

• Hear a short explanation of the format and approximate length.
• Answer open‑ended questions related to their background, skills, location, availability, or other criteria defined in your feedback form.
• Can ask the AI interviewer to clarify questions (for example, 'Can you repeat that?').

If Evaluate English language proficiency is enabled:

• The interview runs as usual, but the system also assesses spoken English using a CEFR‑based rating in the background.
• The rating and details appear later in the recruiter feedback form, not to the candidate.

Candidates don’t see AI scores during the interview.

Coding interviews

In a coding interview, candidates solve programming problems in an integrated coding environment designed to validate logic and correctness automatically.

Screen layout

The main coding interview screen includes:

• A problem panel that shows the full problem description, constraints, and examples.
• An in‑browser code editor where candidates write their solution. The editor automatically loads boilerplate code (a starter template with the correct function signature) based on the selected language, allowing candidates to focus immediately on the logic.
• Options to select from supported languages. While the system supports various languages, core support includes Python, Java, C++, and JavaScript, where type mappings and templates are fully optimized.

Running code and getting feedback

Candidates can interact with the secure execution service to validate their solutions:

• Execute against test cases: By clicking the Run button, candidates execute their code against pre-defined sample test cases to verify correctness.
• Immediate feedback: The system provides instant validation, displaying a Pass/Fail status for each test case.
• Detailed logs: Candidates can view the Standard Output (stdout) for debugging and receive detailed Standard Error (stderr) logs if compilation or runtime errors occur,.
• Sandboxed environment: All code runs in a secure, isolated sandbox to ensure safety and prevent malicious activity.
• Rate limiting: To ensure system stability, code execution requests are rate-limited (approximately 5 requests per minute per candidate).

Anti‑cheating

To maintain fairness, candidates may notice that:

• Copy‑paste into the code editor is blocked.
• Screen sharing for the coding area is disabled.
• Additional proctoring checks can flag suspicious patterns, such as frequent tab switching.

Focus mode

Focus mode allows candidates to code without the AI talking over them.

• What it does:

  • When focus mode is on, the AI interviewer stops proactively speaking and lets the candidate think and code quietly.
  • Candidates can still run code and ask for help manually.

• How to enable it:

  • Candidates click Focus mode from the interview toolbar.
  • The UI highlights focus mode with an icon and color changes on the button and agent frame.

• What the AI agent says: Early in the interview, the AI briefly explains that focus mode is available and what it does.

• How it affects scoring:

  • Transcript segments recorded while focus mode is on are filtered out when generating AI feedback, so silent thinking time isn't penalized.
  • The transcript still shows Focus mode enabled/disabled markers for transparency.

• Turning it off:

  • Candidates click the Focus mode button again to turn it off.
  • If they end or rejoin the interview while focus mode is on, the system automatically inserts a final 'focus mode disabled' marker.

Functional interviews

In a functional interview, the AI focuses on role‑relevant competencies such as problem solving, communication, stakeholder management, or leadership.

Question style

Candidates typically experience:

• Behavioral or situational questions
• Follow‑up questions when their initial answer is incomplete or missing important dimensions (for example, impact or decision‑making).

Candidate expectations

The AI interviewer makes it clear that:

• The candidate should answer with specific examples from their experience.
• It may ask for more detail if needed to complete the evaluation.

As with other types, the candidate doesn't see competency ratings or AI scores during the interview; those appear later in the recruiter feedback form.

After the interview

When the interview ends, the AI interviewer:

• Thanks the candidate for their time.
• May ask if they have any questions and answer within the scope of the configured knowledge base (for example, basic company or role FAQs).

Behind the scenes, the system:

• Generates a transcript and recording (audio or audio+video, depending on admin settings).
• Marks the interview as completed or incomplete, depending on whether it reached the configured completion threshold; scores are only generated for completed sessions.

The candidate doesn’t see the internal summary, ratings, or scores. Those appear only to recruiters and hiring managers in the feedback form and pipeline views.