Safee-Naaz Siddiqi | Senior Associate | Knowledge Management and Competition Law Practices | Cliffe Dekker Hofmeyr (CDH) | mail me |
Artificial Intelligence (AI) tools are rapidly transforming legal workflows. They now accelerate research, streamline contract analysis and enhance eDiscovery. But while these tools offer impressive speed and pattern recognition, they still cannot replace the strategic judgement, foresight and nuance that define quality legal practice.
I recently referenced a pioneering study by Vals that helps clarify this point. The Vals Legal AI Report (VLAIR), which benchmarks legal AI tools like Harvey and CoCounsel against human lawyers on real-world tasks, provides important insight.
AI tools outperform lawyers in certain tasks
The results show that AI tools consistently outperform lawyers on tasks such as document Q&A, summarisation and transcript analysis. However, the study also confirms that humans still have the edge when it comes to redlining and deeper legal research.
In addition, the risks associated with AI use require human oversight in every task, even those that AI tools appear to complete more efficiently. This is a critical reminder that balancing AI and human strategy is not optional; it is essential.
In this context, AI is pushing us to rethink how we approach legal workflows. For example, at our organisation, we are asking tough but necessary questions. What should lawyers be spending their time on? Where should junior training now focus, given that certain types of work can be automated? And how do we ensure that lawyers are trained to use AI both responsibly and effectively?
I have seen how AI tools have already improved efficiency in time-intensive areas. In the past, we would spend hours wading through databases in search of relevant case law. Now, AI-powered legal research platforms allow us to ask questions in plain English and get targeted results far more quickly than traditional methods.
AI is not perfect
Even so, these tools are not perfect. The outputs still require careful review and verification by legal professionals to ensure accuracy and relevance. When used correctly as a starting point for research, they help us work faster and more precisely.
So, it is not just about speed. Used correctly, AI makes us sharper. The tool performs the initial sweep, which allows lawyers like me to spend more time thinking critically about the results. The three, four, or even ten hours I would have spent searching are now invested in deeper analysis and better strategy. That is another example of balancing AI and human strategy in legal practice.
Similar benefits are emerging in litigation. AI is helping to surface the most relevant threads in eDiscovery. It also maps sentiment and patterns that would be difficult to detect manually. Still, these tools do not eliminate the need for human legal professionals. If anything, they underscore just how indispensable human insight truly is.
The irreplaceable role of human strategy in legal practice
It is important to clear up a major misconception. Some believe that adopting legal tech means lawyers need to become coders or data scientists. That is not the case. At the same time, it is not enough to apply AI tools blindly, without understanding how they work.
It is not just about knowing which button to press. It is about exercising sound judgment, understanding what the tool is optimised for, where it falls short, and how to make informed decisions around its use.
One of the biggest shifts I see is in how we value lawyers. It is no longer about the volume of output. It is about the quality of strategic thinking. If all of us have access to the same AI tool, and it only takes five minutes to prepare a discovery affidavit, what will separate a great lawyer from an average one?
The difference lies in how far ahead we think. The great lawyer is the one who uses that extra time to plan 10, 20, or even 30 steps ahead. That is where we create real value. AI tends to think literally. We must think laterally. That is the heart of balancing AI and human strategy. This is especially important in high-stakes matters.
Tactical legal decisions require human thinking
Drawing on my litigation experience, I know that when I receive a letter of demand, I am not just thinking about how to respond. I am considering how my response will look if it ends up in the Supreme Court of Appeal.
AI may offer technically correct answers. But it cannot think strategically about leverage, timing, or positioning. It cannot assess what matters most to your opponent or adjust your strategy accordingly. These are broader tactical decisions that still require human thinking.
That is why our training strategy does not just focus on how to use AI tools. It also teaches lawyers how to interrogate and challenge their outputs. We train our professionals to look at an AI-generated clause and say, “This is why the difference is better”, or to reject it entirely when necessary.
In conclusion
Data literacy is also becoming essential. This does not mean coding. It means understanding the basis for an AI tool’s recommendation, knowing how confident to be in its outputs, and recognising when to override it.
Ultimately, AI is just a tool. Holding a katana will not make me a samurai. What still separates a great lawyer from an average one is the ability to think clearly, to strategise effectively, and to anticipate what is coming next.
