How We Use AI

Back in the 1990s, our senior partner Oleg Buzko kept the main legal codes up to date literally by hand. He would cut legislative amendments out of the official newspaper and paste them into the text. At the time, pagers were considered a modern way to stay in touch with clients.

The legal profession has come a long way since then. The introduction of digital research platforms such as Lexis and Westlaw, the widespread adoption of personal computers, Microsoft Word, email, digital signatures, and electronic court records have all transformed how lawyers work. Each development has reshaped the tools and processes used in legal practice.

The release of ChatGPT in November 2022 marked another major shift, ushering in the widespread use of artificial intelligence (AI) in day-to-day professional work. Within just a few years, AI tools became familiar to almost everyone.

Legal work has proven especially compatible with modern tools built on large language models (LLMs). Research, data analysis, and document drafting are core components of legal practice, and these technologies are well suited to supporting those tasks.

Early LLMs were far from perfect. They lacked internet search capabilities and often produced unreliable results. But the progress made since GPT-3.5 is substantial. Today’s leading models are powerful enough that no serious discussion of the future of legal services can ignore their impact.

For firms committed to excellence, embracing these technologies is no longer optional. It is a natural extension of our commitment to efficiency, quality, and client service. At Buzko Krasnov, we integrate AI thoughtfully and responsibly into our work.

What follows is our current approach to using AI in client matters. It will continue to evolve. This description is current as of December 1, 2025.

1. Access to Leading Models

All of our attorneys have direct access to the most advanced LLMs available today. These are proprietary, closed-source models developed by U.S. technology companies.

2. Training and Professional Development

Technology is only useful when people know how to use it. We invest in training through internal workshops, external courses, and regular knowledge-sharing within the firm to ensure our team uses these tools effectively and responsibly.

3. Responsible Use and Restrictions

We do not use AI tools on highly sensitive client matters, including issues involving national security or a client’s trade secrets.

Before any information is entered into an AI system, we confirm that it will not be used to train or improve the model.

4. Ongoing Innovation and Testing

Through our international legal platform Skala.io, we launched a chatbot that answers users’ legal questions. We have also tested many of the leading AI legal tools on the market, including Harvey and Spellbook. Low-risk, real-world experimentation allows us to evaluate new technologies and refine our internal practices.

AI is not a marketing slogan or an end in itself. It is a tool that helps us deliver efficient, high-quality work. At the same time, AI cannot replicate human judgment, creativity, or the ability to read nuance.

Understanding a judge’s non-verbal cues, negotiating firmly, developing strategy, exercising sound professional judgment, and building trust with clients will remain the work of human lawyers for the foreseeable future.

Will that always be true? Probably not. It is entirely possible that in 15 to 20 years, the next generation of Buzko Krasnov partners will look back with a smile at how we learned to use neural networks, just as we now smile at memories of pagers and newspaper clippings.