For many law firms, AI is still discussed as something that is coming. At Poelmann van den Broek, the conversation has already moved beyond that point. The question is no longer whether AI will change legal work, but how a firm should adapt before the market forces it to.
That mindset explains why the firm appointed Joost van Dongen as Head of AI in 2025. Not an external consultant. Not someone from IT. A practicing lawyer.
The choice says a lot about how Poelmann van den Broek views innovation. AI, in their eyes, is not a side project owned by technology teams. It is something that touches the core of legal practice itself.
“You can’t approach this like a software implementation,” Joost says. “It changes how people work, think, and deliver value to clients.”
Joost is in a relatively unusual position for someone leading AI strategy inside a law firm. Alongside his legal background in IT law and commercial contracts, he also spent more than a decade building an e-commerce business. That combination gives him a perspective that is both operational and practical. He understands the technology, but also the realities of client work, internal processes, and the pace at which lawyers can realistically change the way they operate.
That balance is important, because despite all the attention AI receives, Joost does not believe law firms benefit from rushing blindly into adoption.
In fact, one of the topics he speaks about most is the risk of overreliance.
For the first time in history, cognitive work can be outsourced at scale. Research, drafting, summarising, analysing: tasks that once demanded significant time and mental effort can now be completed in seconds. The productivity gains are obvious. But so are some of the risks.
“We’re already seeing lawyers make serious mistakes because they rely on AI without understanding what it’s actually doing,” he says.
That concern has become increasingly visible across the legal industry. Cases involving hallucinated case law and fabricated citations are no longer isolated incidents. But for Joost, the larger issue is more subtle. It is the gradual erosion of expertise when people stop exercising certain skills themselves.
The concept is known as deskilling.
Joost often uses a simple comparison: navigation apps. Most people today are less capable of navigating independently because technology continuously does the work for them. In everyday life, that may not matter much. In professions built around analysis, reasoning, and judgement, it becomes more significant.
“AI removes friction,” he says. “But friction is often how you learn.”
That philosophy strongly influences the way Poelmann van den Broek approaches AI internally. The goal is not simply to increase usage. The goal is to develop people who understand both the strengths and limitations of the technology.
That requires a different kind of organisational culture than the one many firms are used to.
At PvdB, teams experiment actively with AI tools, but they are also encouraged to remain critical throughout the process. Lawyers are expected to challenge outputs, verify information, and continue developing their own reasoning rather than outsourcing it completely.
Joost believes the best use of AI starts before the prompt itself.
Instead of immediately asking AI for answers, he encourages colleagues to first formulate their own thoughts, assumptions, or direction of analysis. Only then does AI become part of the process. Not as a replacement for thinking, but as an amplifier of it.
That distinction matters because AI-generated work can quickly become predictable. Similar prompts often produce similar outcomes. Efficient, perhaps, but not necessarily insightful. And in complex legal work, originality and nuance are often where the real value lies.
At the same time, Poelmann van den Broek is clear about the opportunities AI creates.
Routine tasks can increasingly be handled faster and more efficiently. Internal knowledge becomes easier to access. Turnaround times improve. Alternative pricing models become more realistic. Some teams within the firm are already experimenting with fixed-fee structures supported by AI-enabled workflows. Others are building internal prompt libraries so lawyers can learn directly from one another’s use cases and best practices.
The broader ambition behind all of this is not difficult to see: freeing lawyers from repetitive work so they can spend more time on the parts of the job clients genuinely care about.
Strategy. Judgement. Human interaction. Commercial thinking.
“We use AI to create more space for the work that actually requires people,” Joost says.
That human element comes up repeatedly throughout the conversation. While much of the public discussion around AI focuses on automation, Poelmann van den Broek believes the opposite dynamic may emerge inside high-value legal work. As technical tasks become easier to commoditise, personal trust and strategic advisory work become more important.
Clients still want someone who understands their business. Someone who can think with them through uncertainty, risk, and difficult decisions. AI may support that process, but it does not replace the relationship itself.
There is also another reason the firm wants to stay close to these developments internally. Their clients are facing the same questions themselves.
Businesses across industries are trying to understand how AI will affect operations, productivity, governance, and risk. By actively experimenting with these questions inside the firm, PvdB believes its lawyers become better equipped to advise clients experiencing similar transformations.
“We’re not innovating for the sake of innovation,” Joost says. “It has to translate into better outcomes for clients.”
What makes Poelmann van den Broek’s approach interesting is not that the firm claims to have solved AI adoption already. Quite the opposite. There is a certain realism in the way Joost talks about the subject. No inflated promises. No dramatic predictions about lawyers disappearing.
Instead, the focus is on building capability steadily and responsibly.
Training people properly. Creating room for experimentation. Staying critical. Protecting professional standards while still embracing technological change.
In many ways, that may turn out to be the real challenge for law firms over the coming years. Not access to AI itself, but learning how to integrate it without losing the expertise, judgement, and human qualities that make legal advice valuable in the first place.
