Bloomberg Law
June 7, 2019, 8:01 AM

INSIGHT: Ten Ways Machine Learning Will Transform the Practice of Law

Caroline Sweeney
Caroline Sweeney
Dorsey & Whitney LLC

When I proposed using concept clustering technology several years ago to facilitate document review in a large e-discovery matter, the idea did not go over well. Today, this is standard for document review, and we are seeing increasing consideration of many other solutions powered by machine learning (ML) and process automation.

What seemed far-fetched 10 years ago is today transforming the practice of law and delivering a return on investment for law offices and legal teams. Firms that aren’t adopting or at least exploring the use of these solutions will soon find themselves at a significant competitive disadvantage.

10 Areas of Use

At our firm, we are using ML or exploring its use in the following 10 areas, and we are seeing real or potential cost savings and process improvements for each one.

  1. Litigation. ML is now widely accepted in litigation, as illustrated by the acceptance of predictive coding by Courts in the U.S. and abroad. At our firm, continuous active learning (CAL) has become a standard for document review. We continue to explore how we can extend ML beyond document review to other phases of litigation like evaluating deposition testimony or conducting decision tree analysis for settlement decisions.
  2. M&A. Firms are increasingly adopting ML to support M&A due diligence analysis. Instead of associates spending hours manually reviewing contract clauses, ML tools that offer automated clause extraction can do this analysis faster and far more efficiently.
  3. Investigations. We use CAL to quickly and cost effectively identify key documents and generate timelines of events (e.g., financial trades). We use other analytic tools to identify communication patterns key to an investigation.
  4. Information Governance. ML offers great opportunity in this area by helping to mine information residing in various repositories. This can be helpful in managing content, but also in managing risk within an organization. I expect to see this area continue to adopt ML as a necessary data management tool.
  5. Privacy. Some vendors now specialize in using ML to help organizations identify personally identifiable information (PII) and protected health information (PHI) to support regulatory compliance and prevent accidental production of this information in litigation.
  6. Trademark/brand compliance. We have started adapting our ML tools to help us efficiently review trademark watch notices to identify potential infringement. Although this is new, we are already seeing promising results.
  7. Expert systems. Tools that capture attorney expertise related to topics, such as privacy regulations, allow us to provide clients with a gap assessment of their compliance needs.
  8. Client service. Firms are using ML-powered systems to automate some client services. For example, a firm may offer a portal where clients can answer a series of questions and, using document automation, be provided with a completed document. By adding robotic automation, the document generated by the client can be routed to an attorney for review, then back to the client. Various practices, including corporate and immigration, can benefit from this sort of automation. We also use process automation for operational processes, such as a new matter intake workflow.
  9. Client Collaboration. We are exploring the potential use of ML to aid a client in the legal review of marketing materials for their business. At the recent CLOC conference, one of the keynote speakers mentioned they are also exploring ML for this purpose. Dashboards that leverage analytics to present clients with information on their legal spend are also valuable collaboration tools.
  10. Business Analytics. Write-offs are a fact of life for most firms. We are using ML to analyze our write-offs and understand their causes. We can then use this information to improve business processes and reduce the number and amount of the write-offs.

Best Practices

To support successfully moving forward with any of the above use cases, I recommend the following best practices.

  1. Obtain executive support. Identify appropriate senior sponsors and other champions. Build the business case by laying out how ML-powered tools and services will increase the bottom line. Learn from clients and assess how you can empower your Firm to improve client processes. Get attorneys onboard by involving them in assessing tools and participating in pilots. Take small steps at a time, validate and communicate the benefits.
  2. Understand the market. Identify internal opportunities to increase efficiency, automate commoditized services, and provide better access to attorney expertise. Maintain awareness of new technology but leverage existing tools wherever possible.
  3. Prepare for success. Invest in personnel to focus on the process and project analysis required to develop the ML use cases. If needed, pull in third parties for expertise and services. Educate the organization and dispel the myth that ML is about eliminating jobs. It’s about new business opportunities, collaboration with clients, and replacing rote work with more substantive and strategic tasks.

ML and AI are the future, and any firm that wants to stay competitive should get on board now.

This column does not necessarily reflect the opinion of The Bureau of National Affairs, Inc. or its owners.

Author Information

Caroline Sweeney is the director of knowledge management for Dorsey & Whitney LLP and a faculty member of the Compliance, Governance and Oversight Council.