How a 10-person tax firm built AI 'departments' to save 10 hours per lawyer, per week

Most small law firms accept a trade-off: stay boutique and struggle to match the bandwidth of larger competitors, or scale up and drown in overhead. Sebastian Korts rejected that premise. His 10-person tax and litigation firm, Steuerrecht.com, is now using ChatGPT Business to punch so far above its weight that it has effectively built what he calls ‘virtual departments’ — for marketing, contracts, research, and knowledge management — without hiring a single additional person.

The math is blunt. Korts reports saving 10 hours of work per week, personally. Drafting a court submission against a tax authority, which once consumed a full day of extracting and countering arguments, now takes about ten minutes before final legal review. Responding to lengthy tax office filings, a three-day slog in the past, is now wrapped in a few hours. Even the mundane but unavoidable work of generating customized CEO employment contracts or small company formation documents has been automated with custom GPTs. ‘It’s the most boring work, but it still demands individualization,’ Korts says. ‘Now we can finish it in a few minutes.’

The output isn’t just faster — it’s strategically sharper. When preparing publications, the lawyers feed their position into ChatGPT and ask for counter-arguments, producing what Korts calls ‘a much higher-quality debate than we could produce on our own.’ The firm has also built an internal knowledge base of text modules and contract templates, accessible instantly by connecting their data to the platform.

What’s genuinely novel here is the ‘translation’ layer. Korts doesn’t just mean German to English. He uses ChatGPT to reframe dense legal arguments for entirely different audiences — condensing a 35-page court submission on tax fraud into a one-and-a-half-page board briefing, then rendering it in English for a foreign CFO who needs to grasp the criminal exposure. ‘That kind of communication has to be crystal clear, in their own frame of reference,’ he explains. The GDPR-compliant setup, which doesn’t train on client data, makes the entire operation viable in a profession where confidentiality is non-negotiable. For a small firm competing against institutions with deep benches, it’s an asymmetric advantage that doesn’t require a single new hire.