TL;DR: Answer with specifics, not platitudes. Connect a genuine thread in your own work to what this company actually does (their products, mission, technical bets), and for mission-driven labs, show you have engaged seriously with the mission including its hard parts, not recited the tagline. Generic enthusiasm ("I love AI, you're the best") fails. A specific, honest "here is why this work and this place fit what I want to build" lands.
How to approach it. Take it seriously: at OpenAI and Anthropic, mission and values alignment is an explicit gate that can sink a technically strong candidate, and Google's "Googleyness" round weights culture heavily. Prepare a specific, honest answer grounded in your real motivations and concrete knowledge of the company. Avoid both empty flattery and rehearsed mission-speak.
A strong answer is two threads woven together.
- Why applied AI (your real motivation). What actually draws you to building AI systems that ship and touch users, versus pure research or generic software. Make it true to you: "I like the full loop, taking a model from a notebook to something a customer relies on, and the part I find most engaging is the reliability and evaluation work: making a non-deterministic system trustworthy." Tie it to something you have repeatedly done. Authenticity beats a polished story.
- Why this company (specific, informed). Show you know what this company does and why it fits: their products, their technical bets, the problems they pick. "I want to work on enterprise deployment of frontier models, and your Applied team does exactly that with real customers, and your eval-heavy culture matches how I think about shipping." Reference something real (a product, a paper, the engineering blog), not a generic compliment.
For mission-driven labs (OpenAI, Anthropic), go deeper. They listen for whether you have genuinely engaged with the mission, including its tensions. For a safety-focused lab, show you have actually thought about why AI safety matters and the hard tradeoffs (capability versus caution, the difficulty of alignment), not just "AI should be safe." A thoughtful, even slightly critical, take reads as real; a recited mission statement reads as rehearsed. Honesty includes naming what you find compelling and what you are still working out.
The throughline: specific plus honest. A real connection between your motivations and their actual work, with evidence you did your homework, is the entire signal. The failure mode is generic enthusiasm that would apply to any company.
Key takeaways
- Weave two threads: your genuine pull toward applied AI, and a specific, researched reason for this company.
- Anchor "why this company" to one concrete reference (a product, paper, or public engineering practice), never vague praise.
- At labs, engage with the hard parts of the mission and your honest uncertainty; recitation fails.
- The disqualifier is a sentence that would fit any employer; if it would, cut it.
What interviewers probe next.
- "What specifically about our work?" Have a concrete reference (a product, a research direction, a public engineering practice); vague praise gets exposed here.
- (Labs) "What is the hardest part of the mission, or our biggest risk?" They want genuine engagement and honest uncertainty, not a recitation.
- "Why not a different company doing similar work?" Show you understand what differentiates this one (approach, culture, focus), which proves the choice is considered.
- "What do you want to be doing in a few years?" Connect it to a trajectory this role actually enables.
Common mistakes.
- Generic flattery ("I love AI and you're the leader") that fits any company and signals you did no homework.
- Reciting the company's mission statement back with no personal engagement or nuance.
- Pure compensation or prestige framing with no real interest in the work.
- For labs, treating the mission or safety question as a formality, which they specifically screen for and reject on.