While in New York at the start of the year LexisNexis unveiled some of the work it is doing to enhance its research capability, including the introduction of what is referred to internally as an ‘interactive intelligent agent’ within Lexis Advance. This update falls into the category of ‘ones that got away’ thanks to the chaos of Legaltech New York, but we recently asked Serena Wellen (pictured), the senior director who is leading this project for Lexis, to tell us more. Here is what she has to say.
A new tool from LexisNexis slated for release in the next few months provides a tantalizing glimpse into what could be the future of legal research. This new tool is what company spokespeople refer to as an “interactive intelligent agent” that will soon be integrated into the Lexis Advance legal research solution.
Users of Lexis Advance will be able to activate this research assistant, or bot, by clicking on an icon in Lexis Advance. Doing so will open a window on the right-hand side of the screen, which will show the user’s recent research activity, with each chat window corresponding to a research thread the user has pursued in the past. Research threads are automatically saved, and the user can re-engage with a thread at any time by clicking on the chat window.
So how will legal researchers use the tool?
Currently there are two primary categories of use cases that the bot will eventually address. The first highlights the personalized nature of the tool, focusing on an individual user’s prior activities. Human recall of the details of prior work is generally poor; this is an area where machines have a big advantage over us. The new Lexis research assistant will instantly recognize when current work overlaps research you did months ago and show it to you again, saving you time. The second involves a user needing assistance to get successful research results. In that case, the bot can act like an electronic mentor, guiding you with personalized research recommendations.
Say, for example, you are researching a thorny topic and get interrupted by a call or urgent email and forget where you left off. Clicking on the bot will take you right back to where you were. A few days or weeks later, you might enter a search query that focuses on the same topic. The bot will ask you if you want to revisit or leverage that prior work, and it might even be able to suggest filters to narrow the previous results based on your more recent activities.
Maybe a few months down the road you’ve completely forgotten these interactions. You ask the bot for help understanding a legal topic, and the bot first provides some initial guidance – maybe a definition from Lexis Answers. You ask the bot directly if you’ve researched the topic before, and the bot not only shows you that prior research but also reminds you of specific documents you may have saved from those prior efforts.
The new research assistant will deploy AI technologies like machine learning, deep learning and natural language processing that are rapidly changing the nature of research in the legal landscape. The goal is to transform the research process into a more conversational experience that resembles interactions we have with colleagues who can listen to our questions, provide initial guidance and then refine their answers by asking questions of their own. LexisNexis expects the tool will eventually encompass multiple sub-bots for different kinds of research-related tasks.
A major initiative at LexisNexis in recent years has been to improve machines’ understanding of the language of law, so tools like bots can provide guidance and answer natural language questions, and researchers can reduce their reliance on mastery of specialized keywords. Personalisation is an essential part of the experience: The new Lexis research assistant will remember and eventually “learn” from interactions with individual users over time so that it can provide more precise guidance and produce the desired results more quickly.
With all the hype around AI in the legal space these days, we forget that the most successful applications of AI in law today focus on language questions and deep pattern matching analysis. This is already true in contract review and discovery review, and now in legal research. Machine learning and bot frameworks are helping us create a more human interaction between users and the services they rely on.
The new Lexis research assistant is designed to be unobtrusive, appearing and disappearing as the user’s workflow dictates. The coolness factor isn’t really the point. Instead, the idea is to make the complexities and nuances of the law more accessible and more understandable for everyone.
You can find out more about Serena here: https://www.linkedin.com/in/serenawellen/
We look forward to playing with the bot to see how it fares.