Training lawyers to think

Editorial illustration: AI-generated; the people shown are not actual SD Law staff or clients.

Why judgment, not document production, is the real measure of competence

Training lawyers in an AI-assisted practice must develop legal reasoning as well as drafting skill. A plausible document is not proof of competence. Accountable professional judgment requires assessing the client’s situation, weighing competing considerations and reaching a defensible recommendation.

Historically, candidate lawyers have demonstrated competence partly through drafting: applying precedents, structuring pleadings and producing clear contracts. Those skills still matter. AI may assist with a first draft, but it can also generate errors and omit crucial context. Speed and volume alone are therefore inadequate measures of competence. Trainers must examine the thinking behind the document.

Document production was always a proxy for judgment, not the skill itself

Drafting an effective contract or a pleading has never been a purely mechanical task. It has always involved an act of judgment. The attorney must assess the client’s situation, weigh potentially conflicting considerations, anticipate how a court or counterparty will respond, and reach a defensible conclusion despite incomplete information. Document production was the evidence of that judgment, not the judgment itself. As a measure of competence it stood as a proxy because the two were hard to separate; a lawyer could not produce a good document without first doing the underlying thinking.

AI has changed the nature of document production, but not the importance of the thinking. Technical drafting skill – the ability to apply precedent correctly and express an opinion without ambiguity – is still necessary. However, it is not, on its own, sufficient. Judgment is built through exposure to real problems and close mentorship. It involves watching how a senior colleague weighs a difficult call, being corrected on a strategic point rather than a grammatical one, or defending a recommendation under scrutiny. It is not built by repeating a format.

Why judgment is harder to train, and more necessary, than automation makes it look

The concern is evident in a January 2026 LexisNexis survey of 873 UK legal professionals: 72% identified deep legal reasoning as a skills concern and 69% identified verification and source checking; only 2% thought AI strengthened junior learning. These are respondents’ perceptions, not measured proof of AI-caused skill loss or findings about South African lawyers. They nevertheless raise a useful question for local firms: how are juniors learning to test an answer?

If juniors mainly edit machine-generated drafts, they may miss opportunities to construct an argument themselves. That is a training risk, not an inevitable result of using AI. A firm can preserve those learning opportunities through supervised research, first-principles drafting exercises and discussion of competing recommendations.

This leaves law firms with a practical responsibility: teach lawyers to operate appropriate tools and to reason independently. Tool literacy and professional judgment are complementary skills. A junior should be able to explain why a recommendation is sound even when the first draft came from a template or an AI system.

What deliberate judgment training looks like

Training judgment deliberately requires time and supervision; it does not happen through billable output alone. As AI changes some drafting tasks, firms should make learning opportunities intentional. Useful practices include:

• Structured mentorship, where senior lawyers review the reasoning behind junior work, not only the output, treating a flawed conclusion reached soundly differently from a correct one reached by guesswork. Feedback rewards the reasoning, not only a clean final product
• Early exposure to real cases, so juniors sit in on client strategy discussions rather than encountering a matter for the first time only when it becomes a drafting task
• Deliberate reasoning practice, requiring juniors to articulate and defend a recommendation before it is accepted

None of these practices are in conflict with the use of AI tools. However, they require a firm to be clear with its candidate attorneys – and with itself – about the distinction between using a tool well and reasoning well. They are different skills, only one of which is why the client hired the lawyer in the first place.

A practical reasoning review for junior lawyers

1. Facts: distinguish verified evidence, assumptions and missing information.
2. Authority: read the original statute, rule or judgment; check jurisdiction and whether it remains applicable.
3. Options: compare remedies, procedure, likely costs and practical risks.
4. Recommendation: explain the proposed next step, its limits and the evidence needed before action.

The supervisor should test the reasoning before approving the work. Client confidentiality and appropriate limits on AI use remain part of that review.

Accountability as the goal

Judgment training and accountability are two sides of the same coin. A lawyer trained to articulate and defend their reasoning can be held to it. A lawyer trained only to produce an acceptable document, whether by precedent, template or AI tool, may not be able to explain the position they took; they will only know it resembles previously successful positions.

Ungrounded reasoning can expose a client to avoidable risk. South Africa’s Legal Practice Council Code of Conduct requires competence and professional care. Paragraph 18.3 specifically requires proper control and supervision over staff and offices; it is not a stand-alone AI rule. A responsible practitioner must check the work and be able to explain the advice, regardless of how a draft was produced.

SD Law – grounded in rigour, innovative in thinking

SD Law’s own use of new tools and techniques is built on this foundation. Adopting AI responsibly means using it to identify questions and possible approaches that must be checked against authoritative sources, not to truncate or circumvent the training that teaches junior lawyers how to apply patterns to individual client matters, including the complex, cross-border custody cases we handle that require deep critical thinking and considered judgment. Innovation of this kind strengthens the judgment of the firm’s lawyers; it is not a substitute for it. This is what professionally grounded innovation means in practice, and it is the standard we at SD Law train our lawyers to meet.

Judgment – the real product of a law firm

Clients may need valid, enforceable documents as well as advice. The value lies in both the document and the judgment that determines what it should say, whom it should protect and which risks it should address. A firm that trains for reasoning and accountability, alongside technical drafting, is better placed to deliver that value as tools change.

SD Law can help

At SD Law, we train our lawyers to think, not merely to produce documents. This is central to how we prepare every member of our team to handle your case. If you have a legal matter you would like to discuss, contact Simon on 086 099 5146 or email sdippenaar@sdlaw.co.za for a confidential, personal discussion.

Simon Dippenaar

Founder & Managing Partner, SD Law

Attorney Simon Dippenaar has a BBusSc LLB degree and Professional Diploma in Legal Practice from the University of Cape Town, and is an admitted attorney of the High Court of South Africa. He is the founder and director of private legal practice, Simon Dippenaar & Associates, with offices in Cape Town and Gauteng representing South African and international clients.

Read articles