TL;DR
Recruiters and career coaches converge on a 60-90 second answer (two minutes is the ceiling — people start tuning out a monologue around the 1.5-2.5 minute mark). The structure that consistently outperforms is present-past-future: where you are now, the experience that built you, why this role next. AI is excellent at compressing that into a spoken script and pressure-testing it with follow-ups — but only after you supply the one thing it cannot invent: your through-line.
The task
The interview opens with “tell me about yourself.” You have roughly 90 seconds before a hiring manager starts forming a working impression. You need an answer that lands the role-fit story, stays under two minutes, and survives the two follow-up questions it will inevitably trigger.
One myth to drop first: the much-quoted “they decide in the first 90 seconds” line traces back to small lab studies, not real hiring. The better-supported finding is that roughly 70% of hiring decisions firm up after the first five minutes (The Interview Guys). So this answer is not a verdict — it is the frame everything after it gets judged against. Set the frame well and the rest of the interview reinforces it.
When this is the right job for AI
- You have a working resume and at least one role you actually want.
- You can answer “what’s the through-line of my last 3 jobs?” in one sentence.
- You will read this answer aloud — not memorize it.
If you cannot say the through-line, do that thinking first. AI cannot find it for you.
The structure that holds up
| Beat | What goes here | Time budget |
|---|---|---|
| Present | Current title + one recent, role-relevant win | ~20 sec |
| Past | The 2-3 prior moves that build the through-line, each with one number | ~40 sec |
| Future | Why this company, this role, now — named specifics | ~20 sec |
Most coaches frame it as present → past → future (The Muse). The variant below leads with the past arc and lands on the present + future, which reads well when your story is a clear progression. Either order works; what matters is one number per past beat and a final sentence that could only be said to this company.
What to feed the AI
- Target role + level + company
- The 3-line summary of your last 3 jobs (role, what shipped, the metric)
- The through-line in one sentence (“each role moved me closer to building developer platforms”)
- The 1-2 themes the JD emphasizes (e.g. “0-to-1 product”, “platform”)
Any current model handles this well. As of June 2026, ChatGPT (GPT-5.5, free tier included) and Claude (Sonnet 4.6 on the free plan, Opus 4.7 on Pro) both produce clean spoken scripts. Use the model’s “thinking” mode if it offers one — for short, high-stakes copy the extra reasoning pass tightens word choice.
The prompt
You are writing my "tell me about yourself" answer.
Target: Senior PM at a developer-tools startup, ~80 person company
Last 3 roles:
- Acme: PM on internal eval platform; shipped V1 in 6 months; cut eval cycle from 3 days to 4 hours
- BetaCo: PM on dev productivity team; cut CI flakes 35%; launched changeset tool
- GammaInc: SWE on platform team; built a feature-flag service used by 12 teams
Through-line: each role moved me closer to building developer platforms.
JD emphasizes: 0-to-1 product, platform thinking, working with senior engineers.
Write:
- 90 seconds spoken (~180 words).
- Past → present → why this role.
- One concrete number per past role; no generic claims.
- Land on a sentence that explicitly bridges to "this role at this company".
- Cut every word that doesn’t earn its place.
Then list the two follow-ups an interviewer is most likely to ask about THIS answer.Sample output structure
“I’ve spent the last six years moving from building developer infrastructure as an engineer to shaping it as a PM. At GammaInc I built a feature-flag service used by 12 teams — that’s where I realized I wanted to own the why, not just the how. At BetaCo I led the dev productivity team and cut CI flakes 35% by changing what we measured. Most recently at Acme, I shipped V1 of our internal eval platform in six months, cutting eval cycle time from three days to four hours. The through-line is platforms for developers — and what pulls me to your team is that you’re building exactly that, at the stage where one PM can still own a meaningful slice.”
Likely follow-ups:
- “Walk me through the eval-platform launch — what almost broke it?”
- “Why move from engineering into PM?”
That word count is deliberate: ~180 words read aloud at a normal pace lands right at 75-80 seconds, inside the 60-90 second window. If the draft runs long, cut to 150 words before you cut numbers.
Practice it out loud, not just on the page
A script that reads well can still stumble when spoken. Run a real rehearsal:
- Open the ChatGPT mobile app, paste a prompt that ends with “act as the interviewer, ask me to introduce myself, then ask one follow-up based on my answer,” and tap the headphone (voice) icon.
- Put the phone down and answer out loud — do not read the script verbatim. ChatGPT replies in voice and asks an adaptive follow-up, which is the part you can’t rehearse alone.
- After two or three rounds, switch back to text and ask: “Where did I ramble? Which sentence had no number? What did I say that I couldn’t back up?”
Voice mode is a supplement, not a substitute — it can’t read your body language or nerves, so still run one round with a human. For a deeper drill, see our AI mock interview walkthrough.
How to refine
- Too vague → add to prompt: “no abstract nouns like
growth,impact,valueunless followed by a number.” - Too long → strict cap at 150 words and read it aloud with a timer.
- Doesn’t land on the role → add: “the final sentence must name something the company specifically does.”
- Sounds memorized → add: “use one contraction (I’ve, that’s) per 50 words to keep it spoken, not written.”
Common mistakes
- Reciting the resume in chronological order. The interviewer can read it.
- No numbers — every claim sounds inflated.
- “I’m passionate about” — drop it. Show the through-line; don’t announce it.
- No bridge to the company. The answer should sound impossible to repurpose verbatim for any other interview.
FAQ
- How long should it really be? 60-90 seconds, two minutes maximum. Listeners start tuning out a monologue around the 1.5-2.5 minute mark, so the sweet spot leaves room for the interviewer to ask a follow-up rather than cut you off.
- Memorize or improvise? Practice the through-line and the numbers cold; improvise the connective tissue. A fully memorized answer sounds memorized — interviewers notice.
- Will AI make my answer sound generic? It will if you feed it a generic prompt. Force specifics: one real number per past beat and a closing sentence that names something only this company does. An answer you could reuse verbatim at another company is the tell of an AI-written script.
- What if I have a career gap? Address it in one neutral line, then move on. Do not let it eat your 90 seconds, and don’t apologize for it.
- What if my last role isn’t the relevant one? Lead with the through-line; chronological order can flex. The present-past-future order is a default, not a rule.
- Free ChatGPT or Claude, or do I need to pay? The free tier of either is plenty for this task. ChatGPT Free (GPT-5.5) and Claude Free (Sonnet 4.6) both write and critique a 180-word script fine; paid plans mainly buy you more voice-mode practice rounds before hitting limits.
Related
- AI behavioral interview prep
- AI mock interview
- AI thank-you email after interview
- AI resume bullet rewrite
- Tell Me About Yourself Prompts: 12 AI Drafts of Your Opening Pitch
Tags: #AI writing #Job search #Interview #Self introduction #Behavioral interview