Hello dear visitors,
in this section of my website theailawyer.org you will find exclusively content created by me personally – AI-generated content has no place here.
This creates a space on this site where I can share my personal views and commentary. I also use this section to explain the background and purpose of this project.
Why did I create theailawyer.org?
As a lawyer, my involvement with IT issues was for a long time primarily that of a requirements provider – defining what I needed in terms of legal and compliance applications and presenting those requirements to an IT department or an external vendor. I would then wait (often for weeks) for results. I was generally glad when my computer worked and I could simplify my professional and personal life with a bit of IT support. I had never shown any particular interest in the technical side of things. That world seemed too cumbersome and abstract to me, and I assumed it would stay that way forever.
When I took on my current role in data protection in 2018, I inevitably engaged more deeply with the topic of data and IT systems. But a genuine interest in IT subjects still did not develop.
When OpenAI released ChatGPT based on GPT-3.5 in November 2022, news of the AI revolution reached me too, and I began looking more closely at the topic in early 2023.
As was probably the case for most of us, I was deeply impressed and fascinated by the results these language models were producing. The spark was lit. I found it particularly exciting to think about and discuss how this new technology could be meaningfully integrated into my day-to-day work and into large corporate organisations, and I ran several projects on this with my team.
Then came January 2026: I heard about Claude Code through my information channels and did not hesitate to sign up for the somewhat more expensive account. This experience has fundamentally changed my perspective on the subject. Since mid-January, I have been spending a great deal of time running all kinds of projects with Claude.
The website theailawyer.org is one of the outcomes of my efforts to become productive with Claude Code.
If someone had told me a few weeks ago that I could build and run my own website, I would simply have laughed at them. Today I know that ANYONE (reasonably tech-open and curious) with a computer and internet access can build and run a website.
With this website I therefore pursue two goals: on the one hand, it serves as an experimental playground to discover and demonstrate what is possible with AI in February 2026. On the other hand, I want this website to be a platform for examining the use of AI in a legal context. That aspect has two dimensions for me: 1. How can lawyers use AI directly to work more effectively? 2. How can lawyers use AI to independently develop and operate the tools they need?
My Experimental Playground
This website and its content (with the explicit exception of this section) are 100% AI-generated. I did not write a single word of the content myself, and I did not write a single line of code for it. The website currently visible is the result of many prompts and several conversations with Claude to overcome technical hurdles. I will continuously evolve the website as I develop new ideas and find ways to implement them.
How can others do the same? Simply ask Claude (or any other AI of your choice)!
An Information and Exchange Platform
This website is not intended for AI experts. Rather, I want to publish information and perspectives here on the meaningful and effective use of AI in the legal field and in particular within large corporations.
I am of course aware of the irony of having an AI shed light on the question of what impact the AI revolution will have on the work of lawyers. It will be fascinating to see whether and how the AI assesses its own role, and what future scenarios and visions it presents to us. I want to make clear that the content is generated by an AI. My intervention will initially consist of influencing the strategic direction of the content when I feel that is necessary. I would also intervene if objectively incorrect information were to be presented. When I want to add or change substantive topics, I always craft my prompts so that the AI is guided by verified information and sources and applies academic standards. There should always be evidence and references wherever possible.
My First Experiences
After just the first few hours of my tentative experiments with Claude Code, it was clear to me that the AI revolution has now genuinely arrived and it is not an exaggeration to speak of a revolution. I would even go so far as to call it a genuine liberation. With these new tools it is possible to independently design and deploy applications. There seem to be almost no limits to creativity.
As lawyers, we were always (only) the requirements providers for IT systems. We had to explain to IT colleagues or external service providers which tools, workflows, upload fields and buttons we needed in order to, for example, digitalise a data protection management system or a business partner due diligence process. Weeks later we would see results, and the next release would then be many months away again.
Those days are probably over. I am firmly convinced of that. Compliance and legal applications can be brought to at least an MVP stage with good prompting (Vibe Coding) in a manageable amount of time – and it is actually great fun. In just a few hours I created a family app (shared shopping list, shared calendar, shared expense tracking, chat function) and got it running synchronously on our smartphones. In just a few hours I prompted the basic framework for a data protection management platform. Creating and launching this website was also accomplished in just a few working hours. I have not hit any limits so far, and the AI has made no promises about feasibility that ultimately could not be kept.
That is why I am so full of enthusiasm and drive. With this technology I am reaching an entirely new dimension of effectiveness, productivity and creativity. There seem to be no limits, and opportunities are opening up that I would never have dared to dream of. It feels as though an insurmountable barrier has suddenly fallen.
Perhaps the assessments of Matt Schumer ("Something big is happening") are infused with a generous dose of Bay Area hype. But I do see a substantial core to them.
Also very fascinating are the developments around OpenClaw (https://openclaw.ai); here you can set up an agent that independently handles digital tasks (from managing your inbox to maintaining your social media presence). At present, this approach raises significant security concerns, as the agent must take on extensive permissions of the user in order to carry out these tasks. I will report promptly once I have explored the topic further.
What Does This Mean for Legal and Compliance Departments in Companies?
AI-driven support for lawyers in transactional legal work, in capturing and summarising facts, in researching case law and commentary, or in drafting submissions and opinions should by now have reached in-house counsel too. Numerous providers have positioned themselves and are offering these services, some with slightly different features and capabilities. We will see which providers survive the competition that is now emerging.
Even more interesting from my perspective is the question of the future of developers and vendors of legal and compliance software such as EQS, Proxora or OneTrust. If the trends that are clearly taking shape continue to confirm themselves, compliance and legal departments will soon be able to independently design and operationalise this kind of software. This would bring not only cost advantages for companies. Companies would also be able to build a tool precisely tailored to their needs and could respond very flexibly to requirements for adjustment, without being dependent on slow-moving external forces. They would also have complete sovereignty over their data.
Companies that master this approach will have an enormous competitive advantage.
Naturally, creating a first "theoretical" tool or an MVP is only the beginning, and there are several hurdles to clear before end users within a company can operationally use a newly self-developed tool. The focus will likely be on IT security questions, documentation and IT compliance. But questions such as quality assurance, comprehensive and documented testing, and ongoing maintenance also need to be resolved.
The demands placed on IT departments in supporting such processes are likely to change significantly, and the success or failure of such projects will depend largely on a functioning symbiotic collaboration between requirements providers, internal IT/AI experts and other stakeholders.
What I perceive above all as a major challenge is the breathtaking pace of technological development and the possibilities that come with it. Tasks the AI could not perform – or performed only poorly – six months ago work very well today. Approaches that were state of the art a year ago are already ruthlessly outdated. This rapid pace will accelerate further, and the great skill will be in identifying the truly relevant changes and implementing them accordingly.
What does this mean for large corporations? If the future changes (too) quickly, there is ultimately little choice but to set yourself up as flexibly as possible in order to respond swiftly to new technologies. Large companies are not good at setting themselves up flexibly. Large organisations require operating in fixed processes and structures so that many small transactions always follow the same path and lead to comparable outcomes. Furthermore, in established companies that have grown over decades, we encounter very diversified IT landscapes. This leads to less flexibility and makes connecting the necessary data very resource-intensive.
I do not have a definitive answer to these challenges either. What seems important to me is that companies gather practical experience as quickly as possible and understand the AI revolution as an ongoing process rather than a one-off disruption to be worked through. Another interesting question will be whether it is sufficient to support existing processes with AI, or whether the processes themselves must be adapted to the capabilities of AI in order to harness its full potential.
I am very curious to see how large corporations will face up to this challenge, and I will report back once I have had further experience to share.
Nicolai Kruck, February 2026
