Most workplace decisions are made under pressure. A performance issue has dragged on. A complaint has landed badly. A restructure cannot wait. A manager’s patience is spent, and something has to be done.
The employer rarely lacks a reason to act. The reason is usually the easy part. The hard part comes afterwards, when the decision, the process, the evidence, and the explanation all have to stand up to someone looking for the flaw. Being right about the problem is one thing. Being able to show it — cleanly, fairly, in a way that survives scrutiny — is another.
Lex Praxis exists in the space between those two things. It helps employers, HR professionals, and advisers make workplace decisions that are clear, evidence-based, procedurally sound, and defensible when they are tested.
Most of that work happens early, in the quiet before a dispute, where a clearer decision and a cleaner process stop a problem from hardening into a claim. That is where the difference is usually made. But a matter does not always stay there, and Lex Praxis does not stop at the advice. Where a case reaches the Employment Relations Authority, or the Employment Court, Lex Praxis represents the employer in it — as an employment advocate, with the litigation experience to match. The person who helped shape the decision is the one who argues it.
Four questions
The work turns on four practical questions.
Can the decision survive scrutiny? Employment process defensibility — the discipline of building a decision that holds together when the Authority, or the other side, takes it apart.
Is the technology helping, or quietly undermining you? AI in employment decisions — whether the tools now sitting inside recruitment, performance management, and monitoring are being used in a way that preserves human judgement, privacy, fairness, and accountability, or whether they are eroding all four while no one is watching.
What do the recent decisions actually teach? Case-based practical guidance — not the headline, but the working lesson from real Authority and Court determinations: what went wrong, what mattered, and what an employer should have done differently.
How do we stop this becoming a dispute? Employer-side risk reduction — the ordinary work of keeping ordinary problems from turning into expensive, avoidable claims.
Where the law meets the tool
Lex Praxis starts with the law and works back towards the technology, not the other way round. That order matters. Most of what is sold as “AI for HR” starts with the tool and hopes the law will catch up. Here it is reversed. The employment-law obligation comes first — good faith, a fair process, a defensible reason, the privacy of the people caught up in it — and the technology is measured against it.
That is what stands behind a phrase used here often: AI-augmented, not AI-powered. The practitioner is the moat; the tool is the instrument. AI does a great deal of the fast work — reading, drafting, sorting, comparing — but nothing leaves as advice, and nothing shapes a real decision, until a qualified practitioner has reviewed it and is willing to put his name to it. Speed is the tool’s contribution. Judgement is not for sale, and it is not handed to a model.
It is also why every output is verified rather than trusted. AI is treated the way you would treat a capable but junior colleague: quick, useful, and every so often confident and wrong. Its claims are challenged, checked across more than one model, and traced back to primary sources — the legislation, the case, the determination itself — rather than to a plausible summary of them. It is asked where it is unsure, and the uncertainty is taken seriously. And at the end it passes through an expert filter, because the accountability for what is said was never the software’s to carry. A model cannot be called to account. It cannot explain itself to the person it disadvantaged. Only a person can do that, and only a person can stand behind the outcome.
Employers hear this and, reasonably, worry about time. If everything is checked, doesn’t verification cancel out the speed? It is a fair question, and the answer is no. Verification is not duplication; it is testing. The reading and the drafting are already done. What remains is the sharper, faster task: pressing on the claims that carry weight, anchoring each back to the primary source — the Act, the case, the determination itself — confirming it against a reliable secondary one, and adding the judgement a model does not have. And the comparison that matters is not verified AI against unverified AI. It is verified AI against the old way, where one person did all of it, alone and from scratch. Measured against that, augmentation with verification is faster — and it is the version that still stands when the decision is tested.
The practitioner is the moat; the tool is the instrument.
Helping people use it well
None of this means keeping AI out of the workplace. That argument was lost before it began. The tools are already there — in the recruiter’s shortlist, in the manager’s draft warning letter, in the investigator’s summary of the evidence — often with no policy around them, no record of their use, and no one senior aware it is happening. The question is no longer whether people will use AI at work. It is whether they will use it in a way that can be defended.
So a large part of the work is preparation, not repair. Lex Praxis helps employers see where AI is already being used across the employment relationship, and where that use quietly creates exposure. It helps them put a workable policy around it — not a document that lives in a drawer, but one people can actually follow. It builds the verification habits and the decision records that turn an AI-assisted decision into a defensible one. And it shows the people who do the work — HR teams, managers, advisers — how to use these tools with their eyes open: what to rely on, what to check, what should never go near a public platform, and how to keep human judgement visibly in the loop.
The through-line is plain. The aim is not to make people frightened of the technology, nor credulous about it. It is to leave the human in charge of the decision, and to leave a trail that shows it.
What it is not
Lex Praxis is not a volume template shop, turning out policies by the dozen. It is not a hype-driven AI consultancy promising that the software will do the thinking. And it is not a law firm; where a matter needs a practising lawyer, or runs beyond the employment jurisdiction, Lex Praxis says so rather than stretch past its lane. It sits in the practical space before things go wrong — where better judgement, a cleaner process, and clearer evidence can still change the outcome.
The articles and case notes are written in the same spirit. They are not there to chase a headline. They take a recent decision and lift out the part that matters to an employer: what went wrong, why it went wrong, and what to do differently next time. They are not marketing. They are a demonstration of how the work is approached.
The point
Lex Praxis exists for those who want a workplace decision to be more than instinct, paperwork, or hope.
The tools are faster than they have ever been, and they will keep getting faster. The judgement behind the decision is still human, and so is the responsibility for it.
Lex Praxis advises New Zealand employers on employment law and AI governance. This article is general commentary and does not constitute legal advice. For advice specific to your circumstances, contact us directly.