Hype or transformation? Where the legal market really stands on AI

Some are already letting AI create case files, while others are still wondering whether they’re even allowed to upload client data. The German legal market is caught between early automated processes and fundamental uncertainty – and is thus at the start of a genuine transformation. Julia Schmidt talks to Dr Benedikt Epe Quarch, Co-Director of the German Legal Tech Hub, about how far law firms, legal departments, the judiciary and public administration have actually come, what new law firm models might emerge, and what is still holding back the wider adoption of AI.

AI has been a major talking point in the legal market for some time now. Has it already transformed it fundamentally, or are we still at the beginning?

Epe Quarch: The fundamental change is already underway, but it’s far from having reached everywhere yet. Three levels need to be distinguished here. When it comes to access to justice, citizens can now get a first legal impression much more easily themselves. That’s real progress, especially in straightforward cases. At the same time, though, we’re seeing the flip side too – for example, social courts having to deal with a large number of AI-generated claims. Abroad, meanwhile, the first AI-based legal service offerings are emerging. One example is Garfield.Law in England and Wales, an AI-based law firm that helps companies enforce smaller claims.

The topic has now reached almost every law firm and legal department. But its use is often still selective: you enter something, get a response, and then carry on with your existing process. AI might replace a quick look at a commentary or a precedent book, but it doesn't yet change the actual workflow. Real transformation only begins once AI is firmly embedded in processes – and on that front, most organisations are still at the very start.

The third tier is the state, public administration and judiciary. Early projects are emerging here too. SPARK, for example, is testing how AI can support public administration, including the drafting of official notices. In the judiciary, early solutions are being used for anonymisation, preparing case summaries, and in mass proceedings. Widespread integration, however, is still some way off. So all three tiers have engaged with AI and gathered initial experience – but the extent of transformation varies considerably.

Now is definitely the right time to continue the path of transformation and embed AI firmly into processes – on all three levels.

Real transformation only begins once AI is firmly embedded in your processes.

You just mentioned an AI law firm in England. Can you imagine such a model existing in Germany in future?

Epe Quarch: I do believe so. In regulatory terms, though, Germany isn’t as advanced as England and Wales. There, the Solicitors Regulation Authority has, for the first time, licensed a purely AI-based firm – Garfield.Law – to provide regulated legal services. Using an AI-powered litigation assistant, Garfield helps companies enforce outstanding debts through the small claims procedure. Licensed solicitors remain responsible for the work and the outcomes.

In Germany, we're initially seeing law firms that rely very heavily on AI, with human lawyers reviewing the results and taking responsibility. We call these “AI-first” or “AI-native” law firms. The first developments in this direction are already underway. Established law firms could also set up a second unit to pool the cases they want to handle using AI. Even if only 50 per cent of the work is initially done with AI, cases could still be handled far more efficiently and in greater numbers.

Whether legislation will one day allow legal services in certain areas to be provided entirely by AI remains to be seen. In the US, developments have in some cases gone even further: there is already a model in which professional indemnity insurance covers a law firm’s AI output. Whether such models will also come to Germany naturally depends on regulation and the insurance market. But ideas from abroad often become established here step by step. That is why it pays to look beyond our own borders.

What sets law firms that already use AI extensively apart from those still hesitating?

Epe Quarch: Those who test early find out more quickly which matters and work steps AI genuinely helps with, and can then integrate it purposefully into their own processes. Realising that it offers little benefit in a particular area is also a valuable finding.

Which developments are currently being overestimated, and which perhaps underestimated?

Epe Quarch: In my view, what’s often underestimated is just how much AI can already achieve in knowledge work today. I always say: we need to give it hard problems too. Genuinely difficult tasks, large case files, or questions that involve several steps. Our first instinct is often to assume it won’t manage anyway. But only by actually trying can we truly gauge its potential.

AI is overestimated, on the other hand, when it’s seen as an all-knowing machine that will replace us entirely. I also think it’s unrealistic to imagine a future without lawyers or courts. The law organises human coexistence, which means people will always play a central role in it. Technology isn’t an end in itself – it’s about improving our work and our lives.

“Technology is not an end in itself. It's about improving our work and our lives.”

The legal market is traditionally rather cautious. Why do we find innovation so difficult?

Epe Quarch: In my view, that starts as early as legal training. We learn early on to follow the prevailing opinion and stick to the well-trodden path. That’s usually the safer route in exams. Unsurprisingly, this doesn’t exactly foster a particularly innovative mindset.

On top of that, much in the legal market has proven itself over a long time, and regulation has grown accordingly. That often means there’s no immediate incentive to change anything. But right now, those who try new things have the chance to reposition themselves. That doesn’t mean suddenly having a completely different profession. You remain a lawyer, but use technology differently, perhaps developing new offerings or new business models. We don’t need to throw all the guardrails overboard. But this basic attitude – always preferring to stick to the familiar path – is one we should let go of.

So where are these new approaches already emerging?

Epe Quarch: A good example is the role of Legal AI Engineer. Two years ago, hardly anyone had heard of it. Lawyers in this role work together with AI providers or software companies, contributing their subject-matter expertise so that the tools are tailored as closely as possible to legal ways of working. This new professional profile is now gaining significant traction.

Law firms and legal departments are also thinking more about how young lawyers can learn to work with AI from the outset. It's not perfect everywhere yet, but the first steps are being taken. We see something similar at universities, where the first mandatory courses are emerging. I'd find it even better if, looking ahead, dealing with AI were also reflected in the state examination. If we keep the existing examination system – which is certainly open to debate – this competency should play a role in it too.

Which changes will lawyers feel most keenly over the next five years?

Epe Quarch: I see four changes. The first is greater technical independence. Over the past 20 or 30 years, lawyers have largely had to rely on IT companies for software matters. Today, with the help of AI, I can at least solve smaller everyday problems myself that I would previously have had to commission a company to handle. This can make us more independent and open up new possibilities for legal work.

Secondly, the transformation will broaden. There will be fewer people hesitating and more actually using AI. Clients, too, will increasingly expect law firms to work with these tools, because they use them themselves. That said, it can also mean clients arriving at the firm with an AI-generated opinion that doesn't always stand up professionally. We need to be prepared for that as well.

The third point is the integration of systems. Technologically, a great deal is already possible, but practice management software and file management systems are sometimes closed off. Over the next five years, I expect significantly more openness and interoperability – of course with due regard to data protection and data security.

Fourth, the judiciary and the state will continue to adapt their processes. Once courts, authorities and other law firms are working with such solutions, lawyers will struggle to opt out without falling behind in terms of their own capabilities.

Change will probably take longer within public administration and the judiciary. But in day-to-day law firm life, we’re really talking about one or two years rather than five, aren’t we?

Epe Quarch: Definitely. At the very latest, once law firm software, Outlook, Word and AI solutions can be connected simply, comprehensively and in compliance with data protection and professional conduct rules, there’ll be no stopping it. Technologically, that’s already possible today. It’s mainly a question of the willingness of those involved.

Agents with a high degree of autonomy will then also be able to take on initial work processes, such as opening matters or preparing an initial assessment for clients. This will also change law firm structures. After all, there are many other processes and parties involved around a matter – from the secretariat to trainee lawyers and research assistants through to enforcement. We therefore need to think about how a law firm will be organised in future. And that future is no longer particularly far off.

Which questions are you most frequently asked about the use of AI at present?

Epe Quarch: Most often, the question is how to use AI in a legally sound way, with matter data, and how to connect it with law firm software. Alongside that is the bigger question of where this development is heading. My view is a positive one: AI will make us better and create new professional fields, and it won’t replace everyone.

Many law firms are unsure how to handle the use of AI towards their clients. Would you disclose which tools you use?

Epe Quarch: There’s no blanket ban on using AI. If it’s set up properly, its use can be designed to comply with data protection and professional conduct rules. I also believe many clients already expect lawyers to use AI as a tool in their processes, in finding solutions and in handling matters. Of course, it’s a different matter if a firm works almost entirely with AI. In that case, it should say so clearly.

I'd generally advocate transparency and naming the tools used. This can be set out, for example, in the engagement letter. In my view, explicit consent isn't required in every case. Anyone who wants extra reassurance can still provide for it.

Is there a question you're almost never asked, but wish you were asked more often?

Epe Quarch: Far too rarely do people ask how AI actually works technically. In my talks, I therefore often take time to explain, at least in broad terms, how the technology works with vector spaces and weightings. We don’t need to go into mathematical detail. But if you want to be able to join the conversation about AI, rather than just having a vague negative feeling about it, it’s worth understanding how it actually works.

Reading a few good articles can already put you ahead. To my mind, that also fits the legal mindset of wanting to understand and get to the bottom of systems. Anyone who understands the basics can judge the possibilities and limits of AI more realistically.

So your appeal is to engage more intensively with AI. What role do formats like the German Legal Tech Summit play in that?

Epe Quarch: In-person events carry particular weight, especially in such a heavily digitised working world. They’re where you can exchange views on how others are using AI, what experiences they’re having, and how providers’ solutions actually work in practice.

Advancing the legal market requires direct exchange between legal practice, IT, AI development and business. That is precisely where the value of the German Legal Tech Summit lies.

If you had to commit to an answer: are we still experiencing hype, or is this already genuine transformation?

Epe Quarch: For me, it’s clearly transformation. Of course, there’s still an element of hype, for instance when you look at some stock market valuations. But when I look at what the technology can already do today, and what some players in our market are already doing and achieving with it, I’m firmly on the side of transformation. Across the board, though, we’re still at the very beginning.

How would you know that the German legal market has changed for good?

Epe Quarch: For me, that point is reached once AI is no longer used only selectively but is firmly embedded in the day-to-day processes of law firms, legal departments and the state, with the systems linked to one another. Once tasks that people currently carry out are taken over or prepared by AI in future, it’s not just a single work output that changes, but the way we work as a whole.

A further step would be law firms openly stating that they handle a very large share of their work with AI and can therefore offer more affordable services or different pricing models. What matters first, though, is that AI becomes embedded across processes more broadly.

Many thanks for the conversation, Dr Benedikt Epe Quarch.

Secure your ticket for the Summit →

Media partner of the German Legal Tech Summit 2026
Deutscher Anwaltverlag & Institut der Anwaltschaft publishes Anwaltspraxis Magazin, in which this interview first appeared, and supports the Summit as media partner. More about the media partner

This interview first appeared in Anwaltspraxis Magazin, published by Deutscher AnwaltVerlag (22 October 2026). The interview was conducted by Julia Schmidt. Reproduced with kind permission.