AI Recruiter First Round: How to Talk to It

You send a resume, and instead of a recruiter's email you get a message: "Hi, I'm Ava, AI assistant at Company X. Let's chat about your candidacy." And you freeze for a second. It's a bot. But how you handle this conversation decides whether you ever talk to a human recruiter at all.

AI screeners are everywhere now. Ashby, Paradox (Olivia), Interviewing.io, HireVue, Gem, plus a dozen white-label versions baked into ATS platforms. Big companies run them to filter candidates before a recruiter spends a single minute. It's a gate. And it genuinely decides things.

What an AI screener actually does

Most people think the bot just asks questions and logs answers. It's more layered than that. Depending on the platform, it can match your words against keywords from the job description, measure how long your answers are, flag whether you're specific or vague, and some video versions even analyze tone and pause length.

But the core thing it's looking for is match. Not charisma, not chemistry, not culture fit in any deep sense. Formal match: right experience, right stack, right location or work authorization, right salary expectations. It compares your answers against a checklist the recruiter filled in before launch.

• Years of experience: there's usually a minimum, and the bot checks whether you stated it
• Specific tools or technologies: if the JD says Kubernetes and you never mention it, your score drops
• Work authorization and location: often the very first question, and the answer determines whether there's a next one
• Salary expectations: some bots ask this upfront to filter out out-of-range candidates immediately
• Answer length and clarity: too short, too long, or too vague all count against you

How not to fail the AI screener

The first and most important thing: speak the language of the job description. Sounds obvious, but most candidates don't do this. If the JD says "Node.js", don't say "back-end in JavaScript". If it says "stakeholder management", use exactly those words in your answer. Most bots match vocabulary, not synonyms.

Second: be specific. An AI screener can't enjoy a good story. It won't appreciate that you're "passionate about technology" or a "fast learner". But it will register that you named a specific stack, years of experience, team size, or project outcome. Facts beat adjectives in any automated screening.

Third: don't overrun or cut short. The sweet spot for a single answer is 60-120 seconds, or 3-5 sentences in text format. Shorter reads as evasion. Longer and the bot may flag it as filler.

Common questions and what they're really asking

AI screeners have their favorites. Here's what they usually mean, and how to answer without getting cut at step one.

1. 1"Tell me about yourself" - the bot wants structure: role, years of experience, stack or domain. Give it that in the first sentence. Don't start with your childhood or how you got into tech.
2. 2"Why are you interested in this role?" - the bot checks whether you read the JD. Mention one specific detail from it. One sentence is enough.
3. 3"What are your salary expectations?" - name a specific number or a tight range. "Open to discussion" sounds like no answer to a bot and may zero out your candidacy right there.
4. 4"Describe a project where you used [technology]" - this is where the STAR method actually helps. Brief situation, task, action, result. Trackr's AI coach can help you rehearse these answers before you need them.
5. 5"Do you have authorization to work in [country]?" - answer directly and precisely. If it's complicated (TRC, displaced person, sponsorship needed), explain briefly without over-explaining.

A note on video screening

HireVue, Spark Hire, Willo and similar platforms record your video answer, and then AI analyzes both content and delivery. A few things genuinely matter here.

• Look into the camera, not at your own face on screen. It reads as eye contact, and some models actually measure this.
• Speak at a moderate pace. A nervous rush reads as anxiety, even if you just think fast.
• Lighting and background matter less than you'd think, but total chaos behind you or a dark silhouette still sends a bad signal.
• Don't read from a script. Some systems detect irregular eye movement and dock your score. Instead, write bullet-point notes and glance at them once before answering.

When and how to ask for a human

Look, there are situations where an AI screener genuinely can't handle your case properly. And it's okay to ask for escalation. The only question is how to do it without looking like you're trying to game the system.

If the screening is text-based or chat-bot, most platforms have commands like "speak to a recruiter", "escalate", or just "human". Try writing something like: "My situation is slightly more nuanced than standard answers allow - could I speak with a person?" Honest and not arrogant.

When you actually need this: if your work authorization is non-standard (displaced person, specific visa, FOP plus relocation), if your CV has a 2022 gap that can't be explained in one sentence, or if the role description is ambiguous and you want clarity before answering. In these cases it's better to ask for a human than to give an inaccurate answer to a bot and get auto-filtered out. For the record, Trackr's AI coach helps you prepare explanations for these edge cases ahead of time.

After the screening: what comes next

If you pass, you usually get an automated message and an invite to the next stage. If not, also an automated message, just a different one. Sometimes nothing at all. It's not personal, it's just how the pipeline is built.

It matters to track where you pass screenings and where you don't, and what phrasing you used. If you're logging your applications in Trackr, you can add a note to each vacancy covering how the screening went, what was asked, what you said. After 10-15 applications a pattern emerges: where your answers land, and where they consistently don't.

• Log which questions the screener asked and how you answered
• Note the outcome: passed, rejected, or ghosted
• If there's a video recording, watch it the next day. You'll notice things you missed in the moment
• Change one variable at a time: try different salary phrasing, or add specific numbers to your experience answer

An AI screener is ultimately just another interview format. Awkward, a bit impersonal, but predictable. And predictability, honestly, is an advantage. You know the rules of the game. Now you just have to play by them.