For hiring managers
AI interview questions for legal and compliance candidates
Legal and compliance candidates now use AI for contract review, regulation summaries, and policy drafts. The interview has to find out how they control it. AI tools read fast and write with confidence. They also miss carve-outs, misstate clauses, and accept material that should never leave the company. Strong candidates plan for those failures.
The twelve questions below cover the six skills Emeri measures, two questions per skill. Each one comes with what a strong answer shows and what a weak answer shows, so two interviewers can score the same answer the same way. The questions test professional judgment. They do not test knowledge of any one law.
How to use these questions
- Ask for a real example
- Each question works best when the candidate describes work they have done. When an answer stays general, ask what they did the last time, and what happened next.
- Ask every candidate the same core questions
- Pick four to six questions before the interviews start. The same questions make answers comparable across candidates.
- Score the behavior, not the vocabulary
- A candidate who names many tools or uses the right terms has not shown the skill. Score against the strong and weak notes under each question.
- Pair the interview with a short work sample
- If the job uses AI, let the candidate use it in the sample. Ask for the requests they wrote and a note on what they changed. An interview shows how a candidate talks about the work. A sample shows the work.
Asking well
Asking for what you need clearly enough to get useful results. What this skill looks like for legal and compliance professionals.
Question 1
A business partner asks you to "run the new vendor contract through AI." What would your request to the tool say?
A review request with no standard returns a general summary that decides nothing. This question shows whether the candidate writes the request the way a senior lawyer frames a review.
A strong answer
States which party the company is. Names the clauses in scope, such as liability, indemnity, data protection, and termination. Supplies the standard to review against. Asks for each issue with the clause number and the quoted text, in a table a reviewer can check.
A weak answer
Asks the tool to "review the contract and flag any risks." Leaves the tool to guess whose side it is on and what a risk is.
Question 2
What do you tell an AI tool not to do when you ask it for help on a legal or compliance question?
Limits are part of a good legal request. This question tests whether the candidate keeps the legal judgment with a person.
A strong answer
Tells the tool not to reach a view on enforceability or on a contested point. Tells it to rely only on the text supplied and to say when that text does not answer the question. Asks it to mark uncertain points rather than resolve them. Explains why each limit matters.
A weak answer
Sets no limits and edits whatever comes back. Or assumes the tool’s own caution about legal advice is enough of a limit.
Giving good context
Giving AI the background, data, and limits it needs. What this skill looks like for legal and compliance professionals.
Question 3
We keep a negotiation playbook with preferred clauses and fallback positions. How would you use it when you ask AI to review a counterparty’s draft?
A model reviews against an average position unless it is given the company’s position. This question tests whether the candidate supplies the right standard.
A strong answer
Pastes the playbook entries for the clauses in scope, with preferred language, approved fallbacks, and the points that need escalation. States the governing law. Confirms the playbook version is current and leaves out superseded entries.
A weak answer
Tells the tool to use "market standard" terms. Or pastes the whole playbook with no selection and no check on which version is current.
Question 4
You need a first draft of an internal records retention policy. What facts about the company would you give the AI before it writes anything?
A policy built on generic facts does not fit the company that adopts it. This question tests whether the candidate knows which facts shape the obligations.
A strong answer
Lists the countries the company operates in, the kinds of records it holds, the systems that store them, and related policies already in force. Supplies the current rule text rather than relying on what the model remembers. Marks which facts are confirmed and which still need an owner to confirm them.
A weak answer
Asks for "a retention policy for a company like ours." Accepts the retention periods the model proposes without asking where they came from.
Checking the work
Spotting mistakes, checking sources, and knowing what to trust. What this skill looks like for legal and compliance professionals.
Question 5
An AI comparison of a counterparty’s draft against our template reports that the confidentiality clause is unchanged. How would you confirm that?
A comparison that misses a change is harder to catch than one that reports a false change. This question tests whether the candidate checks the rows that look safe.
A strong answer
Runs a document compare or reads both versions of the clause. Checks the defined terms the clause relies on, because a changed definition changes the clause. Spot-checks other rows marked unchanged before trusting the rest of the table.
A weak answer
Trusts the "unchanged" rows and reviews only the flagged ones. Or checks the clause text but not the definitions it uses.
Question 6
An AI tool tells you that a supplier agreement lets us terminate at any time. What do you check before you advise the business on that?
A quoted sentence can be accurate and still mislead without the rest of the agreement. This question tests whether the candidate reads the source in context.
A strong answer
Reads the termination clause in full and the clauses it refers to. Looks for notice periods, conditions, fees, carve-outs, and defined terms. Confirms the tool reviewed the signed version and any amendments. Cites the clause numbers in the advice.
A weak answer
Finds the sentence the tool quoted and stops there. Or accepts the answer because it matches what agreements of that type usually say.
Putting it together
Turning prompts, tools, and steps into a repeatable workflow. What this skill looks like for legal and compliance professionals.
Question 7
How would you set up an AI-assisted intake process for routine NDAs, so business teams get faster answers and legal keeps control?
Routine contract volume is where legal teams most often build repeatable AI processes. This question tests whether the candidate designs one with clear human decision points.
A strong answer
Describes a fixed intake form, saved review instructions tied to the playbook, and a comparison against the standard NDA. Standard terms follow a documented approval path. Departures go to a lawyer. People write the escalation rules. The process keeps a record and someone samples the outcomes on a schedule.
A weak answer
Lets the tool approve NDAs on its own. Or describes one long request that changes each time, with no record of what was reviewed.
Question 8
In a due diligence review with hundreds of documents, how would you use AI and still know what was reviewed?
Diligence findings must be traceable to the documents behind them. This question tests whether the candidate builds a process with coverage and a record.
A strong answer
Groups documents by type and extracts the same fields from each, with clause references. Logs every document processed and every one that failed to process. Reviews flagged items in full and samples unflagged ones. Names the documents that need a lawyer to read them in full.
A weak answer
Uploads the data room and asks for "red flags." Cannot say afterward which documents the tool read or skipped.
Getting real work done
Using AI to finish real work, not just make a first draft. What this skill looks like for legal and compliance professionals.
Question 9
Walk us through a contract, policy, or memo you produced with AI help. Which judgments did you keep for yourself?
In legal work the position and the risk call belong to a person. This question tests whether the candidate can show where their judgment went into the final work.
A strong answer
Names the parts AI handled, such as a first draft, an extraction, or a comparison. Names the decisions they made, such as the legal position, the risk rating, and the final wording. Describes what they checked against the source. Speaks honestly about what went wrong.
A weak answer
Cannot separate their work from the tool’s. Or describes the tool taking the legal position and themselves editing the wording.
Question 10
A sales leader needs a one-page explanation of our data processing terms for customer calls by tomorrow. How would you produce it with AI?
Plain-language explanations must stay accurate as they get simpler. This question tests whether the candidate can deliver on time without losing a qualification that matters.
A strong answer
Gives the tool the actual terms and the audience. Uses the draft as a start and checks each simplified statement against the clause behind it. Keeps the qualifications that change the meaning. Adds what sales must not promise and when to bring legal into the call.
A weak answer
Generates the page and sends it. Or simplifies until a conditional right reads as an unconditional one.
Keeping it safe
Protecting private information and using good judgment. What this skill looks like for legal and compliance professionals.
Question 11
Which legal and compliance materials would you keep out of an AI tool, and how do you decide in a borderline case?
Legal teams hold privileged advice, investigation files, and documents under other parties’ confidentiality terms. This question tests whether the candidate has set those limits before a matter forces the decision.
A strong answer
Names specific categories: privileged advice, investigation files, personal data in employment matters, and documents received under an NDA. Uses only tools the company has approved for that class of material. Checks the counterparty’s confidentiality terms. Asks the privacy or security owner when a case is unclear.
A weak answer
Says "nothing confidential" with no categories. Or assumes an approved company tool is acceptable for every document.
Question 12
A colleague used AI to help draft a submission to a regulator. What would you want done before it is filed?
Errors in a filing carry consequences for the company and for the people who sign it. This question tests whether the candidate applies a higher standard where the stakes are higher.
A strong answer
Verifies every citation, quotation, and factual statement against the source. Checks that the draft holds no confidential material beyond what the filing needs. Asks whether the regulator or company policy sets rules on AI use, and follows them. Gets sign-off from the responsible lawyer and keeps a record of how the draft was made.
A weak answer
Treats it as a routine draft and reviews it for tone. Or assumes no rules on AI use apply without checking.
For what each skill looks like in this role at every band from E1 to E4, read AI skills for legal and compliance professionals. These questions are written for interviews. They are not taken from the Emeri assessment.
Questions
What AI skills should I look for when hiring for legal and compliance roles?
Look for all six skills Emeri measures: asking well, giving good context, checking the work, putting it together into workflows, getting real work done, and keeping it safe. In legal and compliance, give extra weight to checking the work and keeping it safe. Clause summaries, citations, and confidential material are where AI use most often goes wrong.
Should a legal work sample use real contracts or matters?
No. Use a mock agreement or a public document. Real matters may be privileged or confidential, and a candidate should not see them. If the job uses AI, allow it in the sample and ask for the requests the candidate wrote and a note on what they checked.
Are these questions taken from the Emeri exam?
No. These questions are written for interviews and are never exam items. The Emeri assessment uses its own questions and scenarios, graded against the same six skills.
How do I confirm a candidate’s Emeri credential?
Every Emeri credential has a public verification page. It shows the band, the assessment date, and a profile across the six skills. The employers page explains how to read one. No account is needed.
Evidence beyond the interview
Emeri assesses the same six skills with graded work, and each credential has a public verification page. Read how to check a candidate’s credential, or see how teams use Emeri assessments.