Dr. Steffen Bunnenberg, co-founder and managing director of LAWLIFT, was a guest on the “Legal Tech Talk” podcast by KSB INTAX. In conversation with Sebastian Bluhm, he set out where AI reliably helps in day-to-day legal work, where its limits lie, and why the combination of AI and rule-based logic delivers better results than using language models on their own. The interview is available as a podcast on the German Legal Tech Hub.
Bunnenberg describes a market that, after a phase of hype, is moving into a more sober assessment phase. Over the past few years, he says, law firms and legal departments have sometimes made investment decisions without first defining a concrete use case. “Buying AI just because it says AI on the label is not a strategy,” Bunnenberg says. Real efficiency gains only emerge when the use of AI is anchored to a clearly defined problem — for example, drafting first document versions, extracting data at scale, or automatically adapting existing templates.
A central theme of the conversation is the division of labour between AI and rule-based logic. Bunnenberg explains that language models reach their limits with structured calculations, and that rule-based systems work more reliably there. The strength lies in combining both approaches: structured data capture via a questionnaire gives the AI the context it needs to adapt a stored document precisely. The result stays transparent and reviewable for the responsible professional.
And professional oversight is not an optional extra. Anyone who wants to scale AI-generated documents and make them available to staff has to review the result before it goes into operation. Technology speeds up the process; responsibility for the content stays with people.
The interview was recorded on the sidelines of the German Legal Tech Hub, where Bunnenberg also gave a talk on AI and rule-based solutions. An accompanying whitepaper on the areas of application is available from LAWLIFT.
🎧 Listen now: Legal Tech Talk by KSB INTAX
Originally published on German Legal Tech Hub.