The person behind Lex Praxis — and why employment law and AI are a more natural pairing than most people expect.
A manager drafts a warning with a chatbot, sifts a shortlist with a tool, summarises an investigation with an AI that no one checked against the file. The decision still has to be the employer's — factually grounded, procedurally fair, independently reasoned, and able to withstand scrutiny when a grievance lands. Lex Praxis exists to keep it that way.
That discipline has a name: Employment Decision Integrity. It is not AI governance in the abstract, and it is not the auditing of hiring algorithms for bias — a different and more crowded field. It is the everyday use of generative AI inside the employment relationship — the warning letter, the restructure rationale, the disciplinary finding — treated as what it is: an evidentiary problem. Get it wrong and none of the failures look like technology failures. They look like an allegation that was never put, a process that exists only in the letter, a decision that cannot be reconstructed when the Authority asks.
Melt Strydom came to this from more than 20 years in employment law. His career began in South Africa, where he was admitted to the Roll of Advocates of the High Court in 2003 and practised as an advocate — the equivalent, there, of a barrister. It was adversarial work, and squarely in employment: he appeared in the Labour Court and the Labour Appeal Court, chaired disciplinary and grievance investigations, and wrote more legal opinions than he could count. In New Zealand — where he also studied the local law in the areas closest to this work — his practice has been advisory rather than adversarial: he advises employers and acts for them at the Employment Relations Authority, and works through the agreements and disputes that sit behind most workplace problems.
Behind the specialist end of that work sit a postgraduate diploma in advanced labour law and qualification as a Fellow of the Association of Arbitrators (Southern Africa) — which, more than any title, taught him to read a dispute from the decision-maker's side of the table, where an employer's process is really tested. What two decades taught him is unglamorous and decisive: sound judgment, careful documentation, procedural precision, and the ability to explain hard things simply to people under real pressure.
Then he spent a long time learning AI properly — formal study in its law and ethics, and a great deal of hard use — far enough to know where it can be relied on and where it cannot.
The work sits at two moments. Before a decision leaves the building, we stress-test it the way the other side's advocate would attack it — the missing fact, the imported foreign law, the consultation that never happened. After a grievance lands, we reconstruct how the decision was actually made — the chain from human input, through the AI, to the person who signed it off. And increasingly it is not only the decision that must hold up but the trail behind it: where prompts and outputs are relevant to whether an employer genuinely turned its mind to the issues, they can be sought at the Authority, subject to privilege. The record has to show that a person, not a machine, made the call.
Lex Praxis is not a technology vendor selling a tool, and not a law firm adding AI on the side. It is one practitioner working where the problem actually sits — employment law, workplace investigations, and an honest read of what AI can and cannot do. Every output is reviewed and owned by that practitioner. The work is AI-augmented, never AI-powered.