.webp)
Quick summary
The problem: Traditional law firm interviews rely heavily on pedigree, experience, and conversational chemistry without meaningfully testing whether a candidate can do the work.
The shift: As AI takes on more first-pass research, drafting, and document review, lawyers create value by directing the tools, challenging their output, and applying legal analysis to a client’s business reality.
Our approach: General Legal built a practical technical interview - a “LeetCode for Law” - to evaluate client communication, AI-assisted legal work, critical review, and commercial judgment.
As an engineer’s spouse, I have learned more than I ever expected to know about LeetCode.
I have watched my partner spend hours working through algorithms, optimising time complexity, and proving that they can actually write code under pressure.
My BigLaw interview process looked very different.
I spent 45 minutes shooting the breeze about law professors, favourite restaurants, and the time I worked for India’s first unicorn technology startup. Variations of that conversation were repeated five times across approximately 15 firms.
It was vibes, all the way down.
The interviews tested whether I could hold a polished conversation with lawyers from prestigious firms. They did not meaningfully test how I would review a contract, identify a business risk, communicate with a founder, or make a judgment call when the answer was not obvious.
Lateral hiring can have the same weakness.
I once interviewed a senior associate who presented himself as a blockchain expert - this was peak 2021. After two substantive questions, it became clear that his practical knowledge of the industry was much narrower than his résumé suggested.
Credentials and experience matter. But they are not proof that someone can do the job.
When I joined General Legal, we were building a law firm designed to deliver legal services at the speed of modern business. Our clients range from solo founders to public companies. Our attorneys work through Slack, email, and live deal rooms, often moving from a product question to a contract issue to a commercial decision within the same conversation.
We also use AI throughout our legal workflows.
Hiring based primarily on pedigree and interview chemistry was not going to cut it.
We needed to see candidates do the work.
The Historic Absurdity of Legal Hiring
The legal industry’s lack of a meaningful technical interview process has never made much sense to me.
Someone may have spent years reviewing documents at a prestigious firm without developing the judgment needed to independently run a complex matter. You may not discover that until they have already joined, are leading intellectual-property diligence for an IPO, and need someone to explain why change-of-control restrictions matter.
That is a terrifying point in the process to discover a skills gap.
For decades, traditional law firms operated through a pyramid model. They hired large junior classes, assigned them substantial volumes of research, document review, diligence, and first-draft work, and billed clients for the hours required to complete it.
The interview process did not necessarily need to establish immediate practical competence. The first years of practice functioned as an extended, highly paid apprenticeship.
That model is now under pressure.
According to Law360 Pulse’s 2026 AI Survey, 70% of attorneys at law firms reported using AI at least once a week.
AI does not eliminate the need for junior attorneys, nor does it replace legal judgment. It does reduce the time required for some tasks that historically occupied a significant portion of junior associates’ days.
First-pass contract review, document comparison, research organisation, issue identification, and initial drafting can now be completed faster with the right tools and appropriate attorney supervision.
That changes what law firms need to evaluate when hiring.
When technology handles more of the initial production work, firms need evidence that an attorney can:
- Give an AI system clear and legally sound instructions.
- Recognise when its output is incomplete, inaccurate, or commercially unrealistic.
- Identify which issues genuinely matter in the context of the deal.
- Communicate those issues clearly to the client.
- Exercise judgment when the most legally protective answer is not the best business answer.
- Take responsibility for the final work.
A traditional conversational interview reveals very little about those abilities.
What the Job Actually Requires Today
At General Legal, AI does not replace the attorney responsible for the work.
Our systems can accelerate initial research, drafting, issue spotting, document comparison, and contract analysis. The attorney still has to direct the system, review its output, identify what it missed, revise the work, and stand behind the final result.
We are not hiring attorneys to copy AI output into a client deliverable.
We are hiring attorneys who know enough to challenge it.
That requires more than commercial-contracting knowledge. We look for clear, precise instructions grounded in legal reasoning. We expect genuine fluency in the client’s industry and business model. We require the ability to adjust tone, detail, and terminology depending on whether the attorney is communicating with a startup founder, a product team, a colleague, or an enterprise general counsel.
A lot of people can talk the talk.
Far fewer can walk the walk.
The 80/20 Inversion
The Harvard Law School Center on the Legal Profession has described a potential “80/20 inversion” in legal work.
Historically, an attorney might spend approximately 80% of their time collecting and organising information and 20% analysing its implications. AI has the potential to reverse that ratio by reducing the time required for information gathering and increasing the relative importance of analysis, judgment, and communication.
That does not make the lawyer less important.
It changes where the lawyer’s value appears.
When an AI system produces an initial contract review, the attorney still has to determine:
- Which suggested changes are genuinely important.
- Which risks are realistic in this particular deal.
- Whether the proposed language reflects the client’s commercial position.
- What the system failed to identify.
- Which issues belong in the redline.
- Which issues are better addressed in a call.
- What the client needs to understand before making a decision.
- Which points are worth spending negotiation capital on.
This requires fast reading and processing. It requires the ability to tailor a markup to the deal's context rather than mechanically applying an abstract legal standard.
Most importantly, it requires the attorney to look beyond the four corners of the immediate assignment and ask the questions a client eager to close the deal may not be thinking to ask.
Those are not skills a résumé can prove.
Building Our “LeetCode for Law”
Because the traditional interview process could not reliably identify these skills, we built our own.
We do not ask candidates to whiteboard their legal arguments or recite the Rule Against Perpetuities from memory.
We simulate the environment in which they would actually work.
The assessment is based on a controlled scenario - not a live client matter - and evaluates four core competencies.
1. Client Communication
We simulate the kinds of written interactions our attorneys have with clients.
We evaluate how efficiently the candidate learns about the client’s product, commercial position, and actual objective. We look at the questions they ask, the assumptions they challenge, and the way they explain the issue.
A startup founder may want a direct recommendation in plain language. An experienced general counsel may want a more technical explanation of the available options and their consequences.
The candidate should be able to adjust their tone, level of detail, and terminology without sacrificing accuracy.
This is not about sounding casual on Slack.
It is about understanding the audience and communicating in a way that helps the client make a decision.
2. AI Direction
We ask candidates to use an AI-assisted legal workflow to draft, analyse, or revise part of a document.
This is not a test of whether someone knows a collection of supposed “perfect prompts.” Nor are we hiring prompt engineers.
We evaluate whether the candidate can give the system clear instructions grounded in sound legal and commercial reasoning.
Strong direction requires the attorney to understand:
- The client’s objective.
- The purpose of the agreement.
- The parties’ relative leverage.
- The relevant legal and commercial risks.
- The desired outcome.
- The limitations of the tool.
A vague instruction produces a generic result.
A precise instruction requires the attorney to know what they are asking for and why.
3. Critical Review and Revision
Once the AI produces an output, the candidate has to review it.
This is the most important part of the exercise.
AI-generated work can sound polished while missing context, overstating a legal position, proposing commercially unrealistic language, or failing to identify an important issue.
We are not looking for someone who can spot a typo.
We evaluate whether the candidate can:
- Identify errors and unsupported assumptions.
- Recognise missing issues.
- Distinguish material risks from minor drafting preferences.
- Explain why the system’s output falls short.
- Improve the substance rather than merely editing the wording.
- Revise the work to reflect the client’s actual business context.
The AI tool does not score or select the candidate. General Legal attorneys evaluate the candidate’s reasoning, communication, and final work.
4. Strategic Judgment
Finally, we introduce situations in which the most protective legal position may not be the right business decision.
The candidate may need to decide whether to insist on a contractual change, accept a manageable risk, suggest an operational workaround, or raise the issue directly with the client before revising the document.
We want to see whether they can look beyond the clause in front of them and understand the transaction as a whole.
That means asking questions such as:
- How much leverage does the client have?
- Is this risk likely to materialise in practice?
- Would insisting on this position delay the deal unnecessarily?
- Can the risk be managed operationally?
- Does the proposed position conflict with something the client has promised elsewhere?
- Which issues matter enough to escalate?
- Which issues are not worth derailing the transaction over?
The obvious legal answer is not always the best commercial answer.
We hire attorneys who understand the difference.
What the Assessment Changed for Us
I initially assumed that the technical interview would mostly confirm what we had already learned from the candidate’s résumé and earlier conversations.
It did not.
The exercise has repeatedly changed our view of candidates.
Some people without the most prestigious conventional background have performed exceptionally well. They understood the assignment, asked sharp questions, identified the real risks, and produced clear, commercially sensible work.
Other candidates with outstanding résumés struggled to translate abstract legal knowledge into practical advice.
We have also received pushback from candidates who believed their experience should speak for itself.
Experience tells us where someone has worked and which matters they may have encountered.
A work simulation gives us evidence of how they think.
No assessment is perfect. A strong attorney can have a bad day, and a single exercise cannot predict every aspect of future performance. We have almost certainly passed on people who could have been excellent in the role.
The technical interview is therefore not the only factor in our hiring process. We also consider the candidate’s experience, substantive knowledge, communication skills, prior interviews, references where appropriate, and the specific needs of the role.
But the assessment gives us something the traditional process often does not:
Direct evidence of how the candidate approaches the work.
Making the Process Relevant and Fair
A technical interview is only useful when it measures skills that are genuinely connected to the role.
Our process is designed around that principle:
- Candidates are evaluated against the same core competencies.
- The exercise reflects the work attorneys would perform at General Legal.
- It does not contain confidential client information.
- Candidate work is not used as uncompensated work for an active client matter.
- Human attorneys - not an automated system - make the hiring decision.
- We assess the candidate’s reasoning, not merely whether their answer matches a hidden preferred response.
- Reasonable accommodations can be made where needed.
The purpose is not to surprise candidates or create artificial pressure.
It is to give them a realistic opportunity to show us how they work.
Competence Over Pedigree
The legal profession is not disappearing because of AI.
But the work is changing.
Attorneys will spend less time manually collecting and organising information and more time evaluating, prioritising, and communicating what that information means for the client.
The lawyers who thrive will not be those who completely resist AI. Nor will they be the ones who trust it blindly.
They will be the ones who understand what it can accelerate, where it remains unreliable, and when human judgment needs to take over.
That requires:
- Enough technical fluency to properly direct the system.
- Enough legal knowledge to identify weak or incorrect output.
- Enough commercial judgment to improve it.
- Enough discipline to verify the answer.
- Enough communication skill to turn the final analysis into useful client advice.
- Enough professional responsibility to stand behind the work.
Our “LeetCode for Law” is designed to identify exactly those attorneys.
It is rigorous. It is demanding. It is occasionally unpopular.
It does not guarantee that every hiring decision will be perfect. It gives us stronger evidence that a candidate can perform the work our clients require.
When a General Legal attorney enters a deal room, we do not want their confidence to come solely from the prestige of their previous firm.
We want to know that they can do the work.
At least we are not asking them to whiteboard the Rule Against Perpetuities.
Looking for outside counsel that scales like software?
General Legal combines experienced legal judgment with AI-enabled workflows to deliver BigLaw-quality legal work in hours, not days. See how General Legal works →
Interested in joining a firm that values demonstrated competence over traditional pedigree? View our open positions →