Warren E. Agin
A former partner of Swiggart & Agin, LLC, at Elevate Mr. Agin designs machine learning systems, helps firms and law departments use and interpret data, and provides subject matter expertise within the LexPredict data science group.
His background includes certified training in data analytics and machine learning in addition to almost three decades experience in insolvency, business and technology law.
Download Mr. Agin's full CV.
Warren Agin is the founding chair of the American Bar Association's Legal Analytics Committee. He has taught analytic techniques as an adjunct professor at Boston College Law School, and writes and lectures on data and machine learning systems.
Predicting Chapter 11 Bankruptcy Case Outcomes Using the Federal Judicial Center Idb and Ensemble Artificial Intelligence (June 1, 2019). Georgia State University Law Review, Vol. 35, No. 4, 2019 (with Gill Eapen)
Using Machine Learning to Predict Success or Failure in Chapter 13 Bankruptcy Cases, 2018 Ann. Surv. of Bankr.Law 13 (2018 WL 4293106).
Bankruptcy and Intellectual Property Deskbook, ABA 2016.
Certifications and Training
Machine Learning - Andrew Ing, Stanford
Machine Learning: Regression - University of Washington
Machine Learning Foundations: A Case Study Approach - University of Washington
Excel to MySQL: Analytic Techniques for Business Specialization - Duke University
Managing Big Data with MySQL - Duke University
Data Visualization and Communication with Tableau - Duke University
Mastering Data Analysis in Excel - Duke University
Business Metrics for Data-Driven Companies - Duke University
Prior to joining LexPredict, LLC (acquired by Elevate Services in December 2018), Mr. Agin was the principal of Analytic Law, and a member of Swiggart & Agin, LLC, a law firm he co-founded in 1999. Internationally known for his work in technology and insolvency law, Mr. Agin has chaired the American Bar Association's E-commerce and Insolvency Subcommittee, chaired the Business Law Section's Technology Committee, and served on the ABA's Standing Committee on Technology and Information Services as well as the Governing Council for the ABA’s Center for Innovation. He also previously co-chaired the Boston Bar Association's Internet and Computer Law Committee, and served for many years on the steering committee for its Intellectual Property Section.
What is Legal Analytics?
Legal Analytics provides new ways to satisfy client needs. Increases in computing power and better access to business data, especially in the last decade, make available to law firms and legal departments algorithmic based techniques for practicing law. Analytic methods can reduce response times, increase efficiency, and allow firms to provide clients with a more accurate and consistent product.
Legal analytics is the art of using analytical methods based on mathematics to solve legal problems and perform legal tasks. Descriptive analytics uses data to describe present conditions. Descriptive analytics can help us understand billing patterns, case outcomes, and prior performance. Predictive analytics uses statistical methods and existing data to predict future outcomes. Predictive analytics can help us understand potential outcomes and improve decision making. Machine learning, a form of "artificial intelligence," uses algorithmic techniques to manipulate information in a reiterative fashion. Machine learning tools can assist document review, enable automated chatbots, optimize decision making, and help build automated prediction tools.
Legal analytics also includes analytic methods that use this data to guide decision making. Behavioral economic concepts such as loss aversion, anchoring and confirmation bias explain party behavior in legal disputes and negotiations. Behavioral economics provides models that allow us to use data to predict third party behavior (and our own), and design strategies to accommodate human actions. When coupled with data analytics, behavioral economics modeling can provide superior decision structures. Game theory techniques, including backwards induction decision trees, also help predict third-party behavior and assist in decision making.
Even though artificial intelligence is now the legal community's hot new word, in reality adoption rates for machine learning and other data analytic tools remain relatively low - most uses within law firms still involve using traditional products, such as discovery and research tools, incorporating some machine learning techniques. But, as we explain, this is rapidly changing as more firms and legal departments start to incorporate analytic methods and tools, and more legal technology vendors build products and services around data and machine learning.
To compete, law firms and departments will have to understand these new tools and their ability to improve performance, increase efficiency, and predict results. Most lawyers lack the background to understand, assess, and properly use these new tools. The typical senior lawyer, even if he or she received training in math, economics or statistics prior to law school, has not been able to use and grow these skills during a legal career. Some techniques, such as behavioral economics, are relatively new, while tools giving lawyers access to data analytics and machine learning only recently emerged. Still, lawyers can learn what they need to know to adopt to the coming changes.
Whether or not most attorneys have noticed it, legal data use has gone mainstream. Products built on analyzing "big data" are now available that can predict the outcome of lawsuits or even specific motions. Small data applications use the data available inside your firm to improve billing practices, measure performance, identify how to generate value for you and your clients, and drive improved results-based decision making. Data driven methods will give the firms using them clear competitive advantages. And, understanding how to manage the data your firm creates is the first step toward using more sophisticated tools, like machine learning and economic modeling.
The legal industry has lagged behind others in using machine learning or "AI" systems. With recent increases in computing power and improved access to machine learning technology, this is rapidly changing. Machine learning already sees widespread use in TAR discovery systems, with some predicting its use in TAR might eventually be mandatory in some situations to reduce discovery time and costs. In the UK, according to consultant Richard Tromans, nearly all top 30 law firms either use or are piloting some form of AI system. A 2017 report by Altman Weil concluded over half of firms with more than 250 attorneys are investigating AI systems. A Bloomberg study early in 2019 found that over 25% of large legal departments were using legal AI systems. Machine learning can allow computer systems to perform complex tasks, but it also performs simple jobs, such as sorting contracts, speeding document review, and finding drafting errors. A growing number of vendors provide machine learning based solutions for many legal tasks, ranging from simple to complex, but new programming tools also make building custom solutions relatively easy. As more large firms incorporate these systems in their product offerings, allowing them to provide improved service more efficiently, the rest of the legal profession must follow.
What is behavioral economics, and how does it relate to the work we do as lawyers? In short, behavioral economics is the science of how people make decisions. By understanding the techniques people use to make their decisions, including those that cause us to occasionally make bad decisions, we can accomplish two things. We can help other people make better decisions (or perhaps, instead, make the decisions we want them to make). We can also better understand our own decision making processes and, with a more concrete understanding, improve them. Understanding decision making can improve our performance in negotiating deals, structuring contracts, or building compliance systems.
Applying behavioral economics concepts, especially when coupled with data and basic game theory techniques, can provide a substantial advantage in making strategic and tactical decisions.
For law firms, the question isn't whether they will adopt data driven practices and techniques - but when and how. For most firms, the next step is building an understanding of these techniques: what they are, how they can help improve delivery of legal services, and how to incorporate them into existing work flows. In many cases, this will mean starting slow. Training will help the firm develop lawyers and other staff comfortable with data driven problem solving. Smaller projects will teach the firm and its professionals how to use data, acquire and build data driven systems, organize and structure work, and integrate new technologies into the firm culture.