1. Improved workflows thanks to small workflow improvements like summarization
I’ve found that improved legal workflows due to smaller tasks being handled by AI will be the first step. A good example is summarization. Using AI-generated summaries significantly boosts efficiency for legal experts, allowing them to breeze through more use cases while upping the quality of their output and workflow. Plus, limiting the focus of the large language model to specific documents helps to restrict hallucination.
However, what I’ve learned firsthand is that crafting the perfect summary for a specific purpose is where the challenge truly lies. It’s not just about generating a summary; it’s about tailoring it just right for the task at hand. Understanding the purpose behind the summary is crucial. Only then can you set the right cues and prompts to get the most relevant summary from AI. I believe understanding this principle will help many legal professionals to integrate AI seamlessly into their daily workflow.
2. Monumental leap for the information provider sector
Another pivotal advancement in generative AI that has caught my attention involves integrating augmented search capabilities in research databases. This comprises four key elements that significantly enhance the search experience.
First, beauty lies in using natural language rather than restricting search inputs to keywords. As NLP (Natural Language Processing) search is not new, the new breed of vector-based embeddings has created a better natural language search. This means legal professionals can simply describe a case or order to find relevant documents.
Second, employing generative AI to improve search engine capabilities, both in terms of the information found as well as the ranking in the results list, markedly enhances the search engine’s quality. Third, this list can be transformed into various formats, be it summaries, memos, or client-ready letters.
Last, the incorporation of a feedback loop facilitates conversational searches, refining the entire process. Implementing this would not only accelerate the legal professional’s workflow but revolutionize it. Not only would it ensure the rapid retrieval of pertinent cases or other case related information, but there would also be hardly any hallucination from outside the research database, all while ensuring proper referencing to information sources. Hallucination is minimized because we force the LLM’s to only work with a narrow set of trusted content that was selected as a first step in the augmented search process. The ability to effortlessly convert findings into client-ready documents also streamlines the operational flow.
At Wolters Kluwer, we’re dedicated to enhancing our services and databases with augmented search functionalities. Our primary focus lies in refining the search engine – a task that poses the greatest challenge. Revamping the search engine using GenAI would signify a monumental leap in the information provider industry. However, in my view, although the first steps are taken right now, mastering this fully will likely take some time.
3. Some contracts are more easily generated than others
Finally, let’s talk about the pivotal role that generative AI will play in streamlining the automation of document creation within the legal domain, which is crucial for enhancing workflow efficiency. Also here, I expect an evolutionary development. The first step is the automation of simple tasks (the summarization or change of a clause). The second is to gather more enhanced information from contracts and the third step will be combining simple drafting with retrieval to really embed the results into the process to create completely new texts. Those texts can actually spark ideas like brainstorming that humans might not have thought about.
Given the vast array of contracts, we must paint a nuanced picture. More straightforward, mass-produced agreements might soon become AI-generated. For example, experiments using large language models to generate leasing or labor contracts are already underway. However, these contracts may not always be of the highest added value.
Developing solutions for bespoke contracts, like exhaustive M&A agreements with detailed asset descriptions, presents a lengthier challenge. These documents demand individualized attention for each case, making automating their creation a far more intricate task that warrants considerable time and effort.
Collectively leveraging the full potential of AI
In essence, technology and generative AI serve as tools to empower legal professionals to accomplish more with an even higher level of quality. This enables legal professionals to focus on more value adding activities and grow their business. The anticipation of how AI will shape the future excites me. Nevertheless, it’s crucial not to get ahead of ourselves. Advancements in AI are gradual, allowing us to collectively explore and leverage its full potential.
