Prepare for a job interview with AI

A research, story-building, and practice workflow that uses AI to improve preparation without inventing experience or replacing the real conversation.

Format
Research playbook
Evidence base
3 cited sources
Reviewed
August 2026 · 10 min

The short version

Interview preparation is a good AI use case because the work is inspectable. You can compare the job description with your experience, organize evidence, practice answers, and notice where your story is thin. The danger is letting a polished draft outrun what you have actually done.

This playbook treats preparation as a case you are building. AI helps you organize and rehearse it. Your experience, judgment, and honest account of the work remain the evidence.

01What should you collect before using AI?#

Save the job description, your submitted résumé, public company information, interviewer details that were voluntarily published, and six to eight real examples from your work. Use public sources for company research and keep personal or confidential material out of consumer tools unless the account and use are approved.

  • The exact job description as a saved file
  • Your submitted résumé and portfolio
  • The company's product, business model, and recent official news
  • The interviewer's public role and career history
  • Examples of decisions, conflict, learning, failure, and measurable results

02How do you turn the job description into a preparation map?#

Ask AI to extract the recurring outcomes, capabilities, and uncertainties in the description, then map each one to evidence you already have. Mark gaps instead of filling them with plausible language. The map should tell you what to research, which story to practice, and which question to ask the interviewer.

ColumnWhat belongs there
Role needA repeated outcome or responsibility from the description
Your evidenceA specific example you can defend
GapA requirement you have not demonstrated
QuestionWhat you need the interviewer to clarify

Tell the model: 'Use only the material I provide. Do not infer skills, results, or responsibilities. Mark missing evidence as a gap.'

03How should you build interview stories with AI?#

Start from rough notes, not a generated answer. Give AI the situation, your responsibility, the actions you personally took, the result, and what you learned. Ask it to flag missing context and unsupported claims, then edit the story until it sounds like something you would actually say aloud.

  1. Recall

    Write the unpolished facts from memory before asking for help.

  2. Separate

    Distinguish what you did from what the team did.

  3. Challenge

    Ask which claims need a number, example, or clearer causal link.

  4. Compress

    Create a 60-second version and a two-minute version.

  5. Verify

    Check every number, name, and outcome against your record.

04What makes AI practice useful rather than performative?#

Practice aloud with role-specific questions, answer without reading a script, and review the transcript for substance before style. Run several rounds with follow-up questions and ask the system to identify repetition, missing evidence, and answers that do not address the question. Treat automated scores as prompts for reflection, not hiring predictions.

  • Round 1: common questions tied to the role
  • Round 2: follow-ups that test the details in your stories
  • Round 3: questions about gaps, tradeoffs, and mistakes
  • Round 4: your questions for the interviewer
  • Final review: three messages you want the interviewer to remember

05What should stay human during the interview?#

The interview itself should be a present conversation, not a retrieval exercise from a hidden script. Listen, ask for clarification, adapt your example, and admit when you do not know. AI can prepare patterns and evidence; it cannot read the room or take responsibility for the claims you make.

Use AI beforeKeep human during
Organize researchDecide what is relevant in the moment
Generate practice questionsListen and respond to the actual question
Find weak evidenceOwn uncertainty and gaps
Review a transcriptBuild rapport and ask follow-ups

Questions people ask#

Can AI predict the questions I will be asked?
It can generate plausible questions from the role and company context, but it cannot know the interview plan. Use it to broaden preparation, not to memorize a forecast.
Should I upload the job description and résumé?
Usually, if the files contain only information you are comfortable providing under the tool's current data settings. Remove addresses, phone numbers, references, and anything unnecessary for the task.
Is it okay to use AI during a live interview?
Only if the interviewer explicitly permits it. Hidden assistance undermines the purpose of the conversation and may violate the employer's process.
How much practice is enough?
Practice until you can answer the core questions naturally without reading and can handle two follow-ups on each important example. More scripting after that often makes the conversation worse.

Sources#

  1. 01
    Practice with AI interview prep

    LinkedIn

    LinkedIn's official description of role-specific practice from real job descriptions, spoken rehearsal, transcripts, and feedback.

  2. 02
    Projects in ChatGPT

    OpenAI

    Current capabilities for keeping files, instructions, and related conversations together for a repeated project.

  3. 03
    Data Controls FAQ

    OpenAI

    Current controls for model training, chat history, exports, and temporary chats on personal accounts.

Related

The timely half

What I'm seeing this week, by email.

By subscribing you agree to Substack's Terms of Use, their Privacy Policy, and their Information Collection Notice.

Playbooks stay current. The newsletter tells you what changed. Free.

The newsletter

AI, translated into decisions you can take action on.

By subscribing you agree to Substack's Terms of Use, their Privacy Policy, and their Information Collection Notice.