For clients, that often sounds like great news: faster turnaround and improved accuracy. But that is not always the case. As powerful as AI has become, it still cannot replace two crucial elements (among other things): legal judgement and experience.

What AI brings to the table

By now, we all have some idea of how powerful AI tools are. Legal AI tools can process and summarise huge amounts of information in seconds. These tools do this by identifying statistical patterns across large datasets using probabilistic models trained on language and structure. Instead of understanding meaning, these tools predict likely outcomes based on prior examples, enabling them to flag anomalies, extract clauses, and generate drafts with remarkable speed across hundreds of agreements. In DD exercises, AI can help locate change of control triggers, unusual termination mechanics, and missing signatures in minutes, rather than hours.

In essence, AI can potentially assist lawyers in spending less time on administrative tasks and more time on strategy, advising their clients on risk management, negotiation, and deal structuring.

The limits of automation

That said, AI tools do not interact with the law or understand nuance the same way humans do. They might recognise patterns, but their ability to grasp context is limited. Unless properly prompted, AI may not compute why a particular indemnity clause is worded in a certain way and may not necessarily process how small drafting changes could shift liability or trigger knock-on effects across a suite of related agreements. Critically, an AI tool may easily miss the commercial pressure behind a deal or the subtleties in regulatory language that can completely change an outcome.

One must also not forget that AI needs to be fed information through intelligent prompts to be effective, which, in a legal context, requires a trained lawyer with a certain level of experience and expertise. Further, while lawyers (and others) may get better at prompting and feeding an AI tool with all the context relevant to a matter – context which accumulates over multiple late night calls, emails, urgent meetings, client preferences and position, and a deep understanding of the client’s strategic goals and the matter history, are what shape a lawyer’s advice and drafting decisions. Even if AI is provided with every minute contextual cue, its ability to appropriately weigh the nuance and interpret relevance correctly is presently limited. Other than taking an enormous amount of time, the prompting process could also end up undermining the efficiency gains that legal AI is designed to provide.

Importantly, AI does not take responsibility for its work. The final advice still carries the lawyer’s name and, with it, human judgement, experience, and accountability. Clients can sleep peacefully knowing their matters are overseen by a human who stands behind the advice, not by an AI that might hallucinate. In recent months, there have been several global examples of AI tools confidently citing non-existent cases or misinterpreting legal principles. These incidents serve as a reminder that AI can assist a lawyer, but it cannot replace one.

Why the lawyer still matters

At its core, legal advice is about applying the law to the facts on hand through the lens of experience, critical reasoning, and strategic thinking.

AI may be able to produce a 200-page DD report in far less time than a human lawyer, but that report is of little value if it cannot reliably assist the client to negotiate the deal or make informed commercial decisions. Lawyers understand that their clients do not want (or often do not have time) to read a 200-page document. They want (and prefer) concise, actionable insights that highlight risks, deal-breakers, and key points for negotiation. In other words, efficiency is meaningless without relevance and clarity.

When drafting agreements, lawyers know what is market practice, when a risk is acceptable, and when it could derail a deal. They understand the counterparty’s position and the strategy adopted by their legal counsel. Lawyers use that insight to decide when to stand firm in negotiations and when to be flexible to get the deal across the line. AI, on the other hand, can help identify risks or suggest clauses, but it does not yet have the commercial instinct or human judgement needed to balance legal precision with deal-making pragmatism.

Way forward? The best of both worlds

The real opportunity lies in combining legal judgement and experience with AI.

When used properly, AI can:

• increase efficiency, freeing lawyers to focus on strategy, rather than routine or repetitive tasks;
• improve consistency, ensuring documents follow best-practice structures; and
• enhance insight, highlighting trends or risks across large data sets.

When layered with a lawyer’s review, these benefits multiply. Of course, the lawyer remains responsible for ensuring that the output is accurate, tailored, and aligned with clients’ goals and priorities. This integration will allow law firms to deliver high-quality, commercially relevant advice more efficiently than ever before.

Ultimately, the future of legal work is not about replacing lawyers with AI but about empowering them. Firms that use AI responsibly are already seeing how it improves service delivery, but also why human oversight remains non-negotiable and key for clients’ needs.

In short, AI can draft, compare, and summarise, but lawyers bring judgement, negotiation skills, and strategic insight shaped by the specific circumstances of a matter. At the end of the day, good legal advice is not just about churning out documents; it is about achieving the client’s goals and objectives.