How Global Law Firms Are Bridging Cultures With AI

In the legal profession, where cases can turn on the interpretation of a comma, global organizations need the highest-quality tools to communicate across languages, cultures, and markets.

Enter DeepL, a leading global Language AI company offering translation and writing tools used by over 100,000 businesses and governments around the world across multiple industries, including the legal field.

DeepL’s AI tools have become a critical investment for companies today, addressing a variety of communication challenges ranging from internal communications between teams to contract drafting and legal correspondence support.

Here, we talk with Frankie Williams, DeepL’s Chief Legal Officer, who sheds light on how law firms are already using this technology — and how AI is making broader inroads in the industry.

ATL: You’ve been in the innovation and startup world for several years now, leading multiple companies through successful IPOs, but you’ve joined DeepL more recently. What attracted you to this company and role? 

FW: I spent just over a decade advising clients on international M&A transactions, and in 2013 I joined King, the mobile game developer that makes Candy Crush — a huge culture shock after so many years in law firms! We listed on the NYSE in 2014 and were acquired by Activision Blizzard in 2016. I went on to spend over five years at luxury ecommerce platform Farfetch, including working on the group’s restructuring and listing on the NYSE at $5.8 in September 2018. During those years, I thrived in the scaling phase of building teams and structures to mature, which is what attracted me to DeepL. I was also drawn to DeepL by its mission to break down language barriers for businesses, which is particularly tangible for me because we have many clients who are legal teams and law firms.

Much of the conventional wisdom in legal tech is that 2024 has marked a transition from an AI hype cycle to the emergence of concrete use cases. What do you make of this idea? 

This is a trend that is becoming increasingly evident in the broader context of AI adoption among businesses globally, as well as in the legal industry specifically. There has been a growing public interest in AI, particularly in the last year, and I think many legal leaders are now realizing that the real value of AI is in its practical, specific applications that are proven to address real challenges and improve efficiency, access, and decision-making.

Recent research from LexisNexis found that 75% of the legal leaders surveyed have used generative AI tools, either personally or professionally, with many believing generative AI technology will decrease costs and increase revenue. Lawyers (me included!) are certainly becoming more comfortable with the technology; I think the question now isn’t whether we should be using it, but rather, which AI tools will dominate in the legal market.

Can you explain the benefits of AI translation and writing tools for lawyers and legal firms? What challenges do these tools help address in the legal field today?

In today’s interconnected world, accurate communication is more important than ever. DeepL is purpose-built to help businesses address language and communication challenges with our suite of specialized AI translation and writing services.

DeepL addresses several common challenges faced by the legal sector, including understanding information in non-native languages, translating documents accurately, and maintaining high writing quality. Our legal customers use DeepL to translate and process contracts, simplify communication with clients and suppliers, improve internal communication, and support research reports and publications. By enabling more accurate and precise communication with clients in different languages, law firms can improve client service and responsiveness.

For in-house legal teams, DeepL is also used for the creation of communications, including translated compliance documents, policies and other essential materials, aiding understanding and compliance. By automating routine linguistic tasks, in-house teams, like my own, can increase productivity and free up time to focus on strategic initiatives that drive greater business value.

According to our customer research, the vast majority of legal department users say that DeepL speeds up their work and improves writing quality. A 2024 Forrester Consulting study found that one legal firm using DeepL was able to cut the time spent translating a two- to three-page document from 2-3 hours to under 15 minutes, thanks to our automation and document upload features.

Contract drafting and other legal tasks require an extreme precision of language, likely beyond that required by just about any other profession. How does AI address this need for precision in translating legal documents and messages?

In today’s highly litigious world, where every word carries significant weight, it is more critical than ever to ensure precision in translation and writing. In our translation quality blind tests with external professional translators in 2024, DeepL’s translation tools were preferred over competitors at a rate of 1.3x more than Google Translate, 1.7x more than ChatGPT-4, and 2.3x more than Microsoft. In the same testing DeepL translations required 2x fewer edits than Google Translate and 3x fewer edits than ChatGPT-4.

We meet the need for precision through several unique capabilities:

• Advanced, highly specialized large language model: The DeepL platform is built on next-generation LLM technology that is uniquely tuned for language, resulting in more human-like translations and writing for a variety of use cases with a reduced risk of hallucinations and misinformation.

• Proprietary data for specific, contextual understanding: Unlike general purpose models that simply train on the public internet, the DeepL model leverages over seven years of proprietary data specifically tuned for content creation and translation.

• Human model tutoring: With a focus on quality, the DeepL model leverages thousands of hand-picked language experts specifically trained to “tutor” the model to best-in-class translation.

Our products offer extensive customization options. Through features like DeepL glossaries, legal teams can tailor translations by specifying the preferred translations of words and phrases for specific contexts. We recently enhanced our glossary functionality to include the industry’s first automatic glossary generator, which streamlines the translation process for documents containing critical brand terms. Our translations can be adjusted to align with the desired tone of voice, depending on the target audience.

And I understand DeepL offers more than just AI-powered translation tools, like DeepL Write. Can you talk about these other DeepL solutions, and how lawyers and legal firms could integrate them into their workflows? 

In 2017, we launched our flagship product, DeepL Translate, which has become our signature product. However, we continue to invest heavily in research and innovation to build out our capabilities. In April this year, we launched DeepL Write Pro, an AI writing assistant that provides writers with real-time suggestions on phrasing, style, and tone while understanding the context of business communication. DeepL Write Pro acts as a creative assistant to writers in the drafting process, elevating their text with real-time AI-powered suggestions on word choice, phrasing, style (simple, business, academic, or casual), and tone (such as enthusiastic, friendly, confident, or diplomatic), helping to ensure the professional quality of writing, enhancing clarity and consistency.

A lawyer’s work is often text-intensive. I find DeepL Write Pro an invaluable tool that enables me to produce precise, high-quality, and consistent writing while reducing errors and improving drafts. I particularly like the new “Shorten” feature which helps counter my lawyer tendency to be overly verbose to ensure my communications to my internal clients is succinct and to the point.

Of course, security concerns are at the core of using AI in the legal industry. Can you walk us through some of DeepL’s privacy and security safeguards? What makes Deepl unique? 

The protection of our customers’ data is at the heart of DeepL’s mission and product offering. Our customer base is diverse, including businesses in highly regulated sectors such as financial services and the legal industry, and ensuring the confidentiality and security of their data is incredibly important to us. We adhere to the highest enterprise-grade privacy and security standards to protect user data — these include state-of-the-art data encryption and the General Data Protection Regulation (GDPR), ISO 27001, and SOC 2 Type 2 standards. We do not use any text from our subscribers to enhance our AI models, understanding that many of our users translate sensitive information.

Everyone with a smartphone can download Google Translate for free these days, getting basic language translation services. What does a more sophisticated service like DeepL do that Google Translate cannot, particularly for lawyers? What technology powers Deepl’s advantages here? 

While the free general-purpose translation services tools you mentioned may work well enough for casual use cases like talking to friends or reading articles in other languages, DeepL offers specialized AI-based translation and writing tools that are secure, sophisticated, and designed specifically for business and corporate use cases where the stakes are far higher. This differentiation is critical for law firms and legal teams.

Unlike general-purpose AI systems, DeepL’s cutting-edge translation and writing solutions rely on specialized AI models tuned specifically for language — we combine the expertise of world-class AI and ML research and engineering with the knowledge of thousands of leading linguists to uniquely specialize in complex language tasks. As I mentioned earlier, privacy and security are also central to our offering. We adhere to the highest standards and state-of-the-art data encryption, which are always priorities for our legal sector customers.

Now, aside from your own company, what new applications of gen AI do you think will be most transformative for the legal industry in the immediate future? And do you have any long-term predictions? 

I am very excited about the potential of AI to improve the ways that lawyers work. My team is using a chatbot, with great success, to answer routine and repeated questions in relation to our customer contracts and also general legal questions (particularly successfully in fields where there is a large bank of law and guidance, like privacy) — saving my lawyers’ time for more complex and strategic work. The ability to extract large quantities of data by simply uploading documents is also a game changer. We will find considerably more exciting and empowering tasks for junior lawyers than the due diligence review of reams of contracts to which I was subjected in the early 2000s!

I think long-term predictions in a field as fast-moving as AI are inadvisable — when I was a junior lawyer the sorts of tools that are already commonplace would have looked like science fiction!