According to a DAV survey (May 2026), 47 per cent of German law firms already use AI tools on a regular basis. But only 23 per cent have documented that use in writing. This is not a technical problem – it is an organisational one. And that is exactly where firms in Bavaria are starting to act when they decide to take the next step.
If you are a lawyer or firm manager thinking about AI, you usually have the same concern: what happens to client data? Where do the documents end up? Who has access? These are legitimate questions – and they have concrete answers.
A mid-sized firm in Wolfratshausen processes dozens of pleadings every day, coordinates appointments, answers client enquiries by phone and email, and keeps its files in order on top of all that. Much of this happens manually, because no reliable alternative existed until now – or at least none that could be trusted without reservation.
The result: qualified lawyers spend a considerable portion of their time on tasks that do not require a law degree. Document preparation, deadline management, initial client communication – that is valuable time that is missing from real legal work.
On top of that, the Federal Bar Association (BRAK, December 2024) has made clear that providers and operators of AI systems will have to comply with specific transparency obligations from 02.08.2026 onwards. So starting now means not only working more efficiently – it also means being compliant in time.
1. Document Analysis and Pleading Preparation
Searching contracts, filtering out relevant clauses, creating summaries for client meetings – all of this can be reliably automated today. If you rely on local processing, you keep full control of the data.
This is where Ollama comes in: the tool lets you run large language models locally on your own hardware – without a single character leaving the firm. No cloud provider, no data transfer, no grey area under GDPR. For firms working with particularly sensitive matters – criminal law, family law, insolvency – this is not a nice-to-have, it is a prerequisite.
2. Intelligent Document Filing with Full-Text Search
Anyone who has ever spent ten minutes searching for a pleading knows how much time gets lost. With LlamaIndex you can build a so-called RAG architecture (Retrieval-Augmented Generation): the system indexes all firm documents and makes them searchable using natural language. You ask: "Show me all contracts with notice periods of less than 4 weeks from the past year" – and you get the result in seconds.
LlamaIndex is particularly well suited to this use case because it works natively with local models such as Ollama. The data stays on your own server. That is the decisive difference from cloud-based document AI tools, where you can never be entirely sure what happens to your content.
3. Automated Client Communication – Structured, Not Robotic
Initial enquiries, appointment confirmations, reminders about missing documents: these are repetitive tasks that still need to feel personal. With AI agents you can build a communication logic that responds to client enquiries, pre-sorts them and – where necessary – escalates automatically.
Important: the lawyer always retains control over the final outcome. The agent handles the groundwork, not the decision.
4. Deadline Monitoring and Task Automation
Missing a deadline is not an option in a law firm. Yet many firms still work with manual calendars or outdated software systems. With n8n – an open-source automation tool that can be self-hosted – you can build deadline workflows that automatically send reminders, escalate and document. Fully on-premise, fully under your own control.
Compared with cloud-based alternatives such as Zapier, n8n has a clear advantage in the law firm context: because it can be self-hosted, no process data leaves your own infrastructure. For privacy-sensitive industries, that is a significant factor.
5. Client Portal with Structured Document Handover
Instead of sending documents back and forth by email – which is problematic from both a legal and a security standpoint – you can build a simple portal through which clients upload documents, book appointments and check the status of their matter. This reduces follow-up queries and takes a noticeable load off the administrative team.
Our clear view: anyone who promises you that an AI will soon independently produce legal opinions or advise clients in a fully automated way is selling you science fiction as a product. AI is a tool – not a substitute for legal judgment.
The concrete limitations:
This is not an argument against getting started with AI – it is an argument for realistic expectations.
In 8 out of 10 of our consulting projects with service providers and law firms, unstructured document filing was the first bottleneck – not a lack of budget, not a lack of staff. If your documents are not cleanly structured, you cannot put an AI on top and expect meaningful results.
That sounds obvious – but it is not. Because many firms think they first need to find the perfect AI system. In reality, they first need to know where their own documents are.
A concrete example: a firm from Holzkirchen reduced the average search time per file from over 12 minutes to under 90 seconds by introducing a local RAG architecture with LlamaIndex and Ollama – while also improving the findability of older pleadings. The entire process was up and running within four weeks, because the data foundation had been cleaned up together beforehand.
If you want to find out more about how we implement automation solutions for service providers or what AI-powered agent systems look like for structured industries – we are happy to show you in concrete terms.
The next step does not have to be a big project. Often it is enough to start at a single point: the document chaos. If you want to know exactly where your firm stands and which AI solution fits your size, your budget and your data protection requirements, talk to us.
We analyze your processes, identify automation potential and show you which digital systems really make sense.
