Law firms urged to codify judgement as AI spreads

Avvoka said law firms must codify their judgement as artificial intelligence becomes more widely used across the legal sector. Firms that rely on the same underlying AI models as their rivals risk losing differentiation, the legal technology company argued.

The point comes as legal practices expand their use of AI tools for drafting, summarising and research, raising questions about how they preserve their identity, train junior lawyers and keep internal standards current. Avvoka, which raised GBP £14 million earlier this year, has argued that a firm's value still rests on human judgement rather than automated output alone.

Broad access to AI is lowering barriers between firms because many are drawing on similar models and reaching similar results, Avvoka said. In that environment, legal advice risks becoming commoditised unless firms can show how their internal reasoning, preferred drafting positions and client-specific standards differ from those of competitors.

That issue is particularly sensitive in a market where clients have long paid for specialist judgement rather than billed hours. Avvoka's co-founders said firms need to make that judgement more explicit by embedding it in templates, knowledge systems and review processes, while ensuring the source material behind any AI-assisted work remains accurate and controlled.

Training shift

Avvoka also pointed to a change in legal training. Junior lawyers have traditionally developed their judgement through repetitive document work, detailed review by senior colleagues and years of observation within deal teams. AI is now taking over much of the searching, summarising and first-pass drafting that once formed the basis of that apprenticeship.

The shift creates both an opportunity and a risk, according to Avvoka. Younger lawyers could reach partner-level reasoning more quickly if firms build that expertise into internal systems. But firms may also lose a longstanding route through which legal instinct and practical judgement were passed from one generation to the next.

That means law firms need to move from passive training to a more deliberate model, the company argued. Rather than relying on junior staff to absorb know-how through observation, firms may need structured exercises, simulations and guided use of internal standards so legal judgement is taught directly.

This, in turn, gives new importance to knowledge management functions that have often been treated as back-office support. If AI systems are to reflect a firm's actual positions and standards, internal precedents, clause banks and drafting guidance need more active curation and more frequent review.

Living standard

Avvoka also argued that template management can no longer be treated as a one-off exercise. Law firms often spend substantial time agreeing house positions and producing standard forms, only for those documents to date quickly as market terms, regulation or client expectations change.

In transactional practice, even a narrow contract term can lose relevance as market conditions shift. The result is outdated clauses, drafting gaps and extra rework, all of which add cost and can undermine confidence in the advice being given.

AI can help reduce that burden by speeding up template reviews and identifying where a firm's standard position has changed, Avvoka said. But this should be treated as a continuous feedback loop rather than a drafting project left untouched until the next major review.

Avvoka launched a template creation product, Curate, this summer, expanding its focus on how firms create and maintain internal standards. The product sits alongside the company's broader push around document automation and legal drafting processes.

Human oversight

Avvoka's comments also touched on concerns about reliability and accountability in legal AI. Courts and regulators have become more alert to the risk of fabricated citations, invented authorities and other false output generated by AI systems, placing greater emphasis on human review.

That trend does not diminish lawyers' role but makes it more central, Avvoka said. In the company's view, a lawyer's task is to judge whether language is right for a particular client, transaction or firm position, even if software handles much of the preliminary work.

Avvoka also linked that point to emerging efforts by AI providers to mark or trace machine-generated text. If such markers become more common, firms may face added pressure to show where human review took place and how final work reflects professional judgement rather than unexamined machine output.

Avvoka said: "Humans in the loop are the point, and what makes law firms unique because of the judgement they have in setting their services apart from their competitors."