Which tool to choose
You don't need a dedicated app: you need the right AI for two different tasks.
To generate the questions: choose Claude if you want to paste a full resume plus the posting and receive questions tailored to your profile. It handles a lot of text, so you can put together the resume and several position descriptions without trimming them. The free plan covers a normal written preparation session. ChatGPT in the free version does the same job if you don't have extremely long documents to paste.
To inform yourself about the company before formulating the questions: use Perplexity, which cites its sources and gives you recent news (funding rounds, office openings, product news) to build pointed questions on.
To try saying them out loud: Google Interview Warmup, free and with no account. Open the site, choose a job category, and record your answers as you speak. You'll find it at grow.google/interview-warmup.
If you're a beginner and in a hurry, start with ChatGPT or Claude. The other two are finishing touches.
How to do it
The path doesn't change whether you're in a browser or an app. Work in two phases: first you inform yourself, then you generate the questions.
- Open Perplexity and search for the company by name plus the position. Gather 3-4 concrete and recent facts: a new product, an office opening, an award, a growth story.
- Open Claude or ChatGPT. Paste the text of the posting (full copy-paste), your resume, and the facts gathered in step 1.
- Launch the prompt below.
- Read the generated questions and discard those whose answer is already in the posting: asking what's already written has the opposite effect.
- Keep 4-5 questions, not twenty. Learn them in broad strokes, not word for word.
The operating syntax:
I'm preparing for an interview for the position described below.
Help me formulate 8 smart questions to put to the recruiter at the end of the interview.
Divide them into 4 categories: 1) role and daily activities, 2) team and how it fits into the organization, 3) growth and training, 4) company culture and upcoming goals.
For each question add, on a separate line, why it's useful to ask it and what the answer reveals to me.
Avoid questions whose answer is already in the posting. Do not include questions about salary.
Use a professional but natural tone, the way I'd actually speak.
POSTING:
[paste the text of the posting here]
MY RESUME:
[paste your resume here]
RECENT FACTS ABOUT THE COMPANY:
[paste the 3-4 facts gathered here]
Check: if the AI returns questions like "What are the hours?" or "How many vacation days?", the prompt did a poor job or you gave it too little context. Good questions hook onto something specific about the posting or the company.
A rule that matters: don't necessarily wait until the end. If you can, insert your questions during the conversation, hooking them to what the recruiter has just said. The dialogue flows and you show you're listening.
A concrete example
Marta applies as a project manager at a medium-sized software house. She searches the company on Perplexity and discovers they've just opened a second office and launched a product module for logistics. She pastes the posting, resume, and these two facts into Claude with the prompt above.
Claude returns, among others, two strong questions: "I read about the new logistics module: would the project manager in this role be involved on that line or on already established projects?" and "With the opening of the second office, how is the team I'd be working with distributed?".
At the interview Marta inserts them at the right moment. The recruiter notices that she informed herself and that she's already reasoning from inside the company. Here's the point: a question hooked to a real fact says you've prepared, and that weighs more than any memorized answer.
When it does NOT work (and how to fix it)
If the AI churns out generic questions
It happens when you give it too little to chew on. If you paste only "marketing role" without a posting or resume, you get textbook questions. Fix: always paste the full text of the posting and at least one recent fact about the company. It's the context that turns "What are the growth opportunities?" into a question hooked to that company.
If it invents data about the company
Generative AI can "complete" information it doesn't have, giving you plausible but false facts about revenue, offices, or products. Never build a question on a piece of data you haven't verified: citing a nonexistent news item at the interview is worse than silence. Fix: gather the facts about the company on Perplexity, which cites its sources, and check the official website. Use generative AI to shape the questions, not for the research.
If the questions sound fake when you say them
Questions written by an AI are sometimes too formal for how you speak. Fix: try them out loud. Google Interview Warmup records your answers as you speak, counts fillers like "um," and flags whether you hit the key points. If you stumble over a sentence, rewrite it in your own words: it has to sound like yours, not recited.
If you go blank and say "No, no questions"
It's the classic mistake. When they ask whether you have questions, "No, nothing comes to mind" sounds like disinterest. Fix: always bring two questions printed or on your phone. Even just looking at them unblocks you.
A tip from someone who really uses it
Ask the AI to also prepare your closing question, the one you keep for last: "Is there anything in my profile that gives you hesitation and that I can clarify?". It's direct, shows confidence, and gives you the chance to dismantle a doubt before the recruiter leaves with that impression. Have it generated tailored to your resume, so it touches your real weak points and not an abstraction.
And remember why it matters: when they ask whether you have questions, silence is read as poor preparation. Asking even just a couple tells the recruiter that you listened and that the role really interests you.
Frequently asked questions
How many questions should I prepare for the interview?
Generate 8 with the AI, bring 4-5, ask 2-3 live. The high number serves only to give you margin: some answers come up spontaneously during the interview, and those you drop. Better a few targeted questions than an interrogation barrage.
Can I ask about salary among the questions to ask?
Yes, but not as the main question and rarely at the first interview. It's normal for salary to count in the choice, but it shouldn't be the center of the first meeting. Usually the topic comes up in later stages, or the recruiter raises it. Keep the first interview on role, team, and growth.
Which themes should I concentrate the questions on?
On role and prospects. Ask what the daily responsibilities are and how long you're expected to take to reach the goals: you immediately understand what they want from you. Add a question about training, which signals a desire to grow, and one about the company's future.
How much does it cost to use AI to prepare the questions?
Zero, for this task. The free plan of ChatGPT or Claude covers generating and refining the questions; Google Interview Warmup doesn't even ask for an account. The paid versions serve those who practice intensively for weeks, not those preparing for a single interview.
Does the recruiter realize I used AI for the questions?
No, and that's not the point. The AI helps you structure and not forget anything, but the questions have to pass through your filter: you choose them, you rephrase them in your own words, you hook them to what the recruiter says in the moment. The recruiter doesn't perceive "they used AI," they perceive "they prepared." And preparation is exactly what gets rewarded: the tool stays invisible, the attention you show does not.