How to Answer: "Why Do You Want to Work at Harvey?"

Harvey wants to hear why you care about AI for legal work, not AI in general. Harvey builds generative AI products for law firms and professional services firms. Its tools help lawyers research, draft, and review documents. Interviewers report that Harvey looks for people who care about the product and can work with non-engineers, such as lawyers. Your answer should prove both things with one concrete story.

The legal domain shapes everything at Harvey. Lawyers work under strict rules about confidentiality and accuracy. An answer that ignores that reality sounds like an answer for any AI company.

What the Interviewer Is Listening For

A reason tied to legal work. Why do you want AI applied to law, specifically? Strong reasons include the high cost of legal errors, the volume of document work, or a personal experience with legal services. Weak reasons are general excitement about language models.

Respect for accuracy. Legal users cannot accept invented answers. A hallucination is an output where the model states something false as fact. Show that you know this problem and want to work on it.

Comfort with non-engineers. Harvey engineers work closely with lawyers and product staff. Interviewers listen for stories where you learned a domain from experts and changed your design because of it.

Ownership and pace. Harvey is a young company that ships quickly. Candidates who wait for full specifications do poorly. Show a story where you moved with incomplete information.

A Three-Part Structure

Part 1: The domain hook (2 to 3 sentences). Name what pulls you to legal AI. Use something concrete: a document workflow you saw, a legal process you experienced, or a retrieval problem you worked on.

Part 2: Your evidence (3 to 4 sentences). Describe relevant work you owned. Search systems, document pipelines, model evaluation, or any product built with domain experts all fit. Include numbers.

Part 3: The direction (1 to 2 sentences). Say what you would want to build at Harvey and why it matters to legal users.

Sample Answer

"I want to work at Harvey because legal work is the hardest honest test for language models. A wrong citation in a brief has real consequences, so the product must prove every claim. At my current job I built a retrieval system for insurance documents, and I cut wrong-answer rates by half using citation checks. I did it by sitting with claims adjusters weekly and turning their corrections into evaluation data. That taught me to treat domain experts as part of the engineering loop. Harvey works the same way with lawyers, and it serves firms where quality is the whole product. I want to build the retrieval and evaluation layers that make legal AI trustworthy."

This answer names the domain, shows matching work with numbers, and ends with a direction. It would take about 60 seconds to say. Notice that the story does not need to be legal. Any domain with experts and high stakes transfers well. The link between your work and Harvey's problems is what counts.

Common Mistakes

• Generic AI motivation. If your answer fits a consumer chatbot company equally well, it fails. The legal domain is Harvey's identity.
• No mention of accuracy or trust. Skipping the hallucination problem signals that you have not thought about legal users.
• Talking down about lawyers. Harvey's customers are experts. Any hint that legal work is simple or ripe for replacement reads badly.
• No product contact. Read what Harvey publishes about its products before the interview. Quoting a real workflow makes your answer specific.

How to Prepare

• Write the three parts and time them. Keep the answer under 90 seconds. Remove every sentence another candidate could also say.
• Prepare stories for the values Harvey screens. Ownership, ambiguity, and shipping fast come up across the loop, as listed in Top Harvey Behavioral Interview Questions (and How to Answer Them).
• Know the full loop. The rounds are described in What Is the Harvey Interview Process Like? (Round by Round).
• Prepare the design round. Harvey asks document and search design questions, covered in What to Expect in the Harvey System Design Interview.
• Train the storytelling method. Grokking the Behavioral Interview teaches how to build answers from evidence.