Publications

The Magazine of Illinois Lawyers

April 2017Volume 105Number 4Page 22

Thank you for viewing this Illinois Bar Journal article. Please join the ISBA to access all of our IBJ articles and archives.

April 2017 Illinois Bar Journal Cover Image

LawPulse

Robots invade Seyfarth

Seyfarth Shaw is using software robots to do some tasks in one tenth the time it takes a human.

The robots are coming - no wait, they're here. It reads like the tagline from a Hollywood movie trailer, but it may be the new normal for the practice of law. Artificial intelligence and software robots were once the stuff of science fiction. Now, they are helping lawyers improve processes in their offices by automating labor-intensive tasks.

Seyfarth Shaw LLP, founded decades ago in Chicago, has led the legal industry in the field of process mapping. The firm has spun off a separate business, SeyfarthLean Consulting LLC, which assists large law firm as well as legal aid clinics with streamlining their legal processes. See http://www.seyfarth.com/SeyfarthLean.

Many of these approaches involve leveraging technology to eliminate redundancies. On February 7, 2017, Seyfarth entered into an agreement with London-based Blue Prism to use the company's robotic process automation software in the firm's practice areas.

Process mapping is well known to those familiar with Six Sigma, a business process discipline used by tens of thousands of companies. A key concept behind the practice is identifying redundancies and inefficiencies in a business's operations. Removing those inefficiencies frees up attorneys and staff to do other tasks.

For example, a document may go through unnecessary layers of review before it is delivered to a client or filed with the court. Firms with multiple practice areas may have multiple practice management systems, each one requiring independent data entry. (Read more about the topic in the August 2016 IBJ cover story (http://bit.ly/2moQJRi).) Robotic process automation software may be a step towards increased efficiency.

Not all robots are machines - or AI

Not all robots are physical machines. Software robots are built entirely from computer code. They've been in use for decades. One familiar example is how search engines traditionally indexed the content of the Internet. Rather than having a team of people scour the Internet, software robots crawl the web, finding websites and content.

Seyfarth's robots will work differently, but the concept is the same. According to Blue Prism's press release, its software robots are designed to "automate high-risk, manual, rules-based, and repetitive tasks" to "radically improve[] agility, efficiency, accuracy and compliance" (http://bit.ly/2kmOEHY).

J. Stephen Poor, Seyfarth's chair emeritus, says that the firm's first robots will be used to automate processes. Specifically, he says that Seyfarth has many practice management systems that are specific to its various practice areas. Some of those practice areas use legacy systems.

He points to the firm's immigration practice as an example. Information that was input into the firm's computer systems would then have to be moved to three other systems. Having a staff member manually perform this task was taking 30 to 45 minutes. Robotic process management has brought that time down to four minutes. Poor stresses that software robots are not the same as cognitive computing, which is what is commonly considered artificial intelligence.

Poor says that software robots allow the firm to "take the robot out of the human." They don't automate a person's job; they automate tasks. The robots work on a set of if/then rules. By using them, the firm frees up its staff to do other tasks. This creates a benefit to the firm and its clients, particularly because clients are billed for four minutes of work, not 45.

Trusting the robots

On the other hand, cognitive computing is beginning to be used to analyze commercial loan agreements. Bloomberg reports that JPMorgan Chase & Co. has rolled out an AI referred to as COIN, or contract intelligence. Hugh Son, JPMorgan Software Does in Seconds What Took Lawyers 360,000 Hours, Bloomberg Markets (Feb. 27, 2017), http://bloom.bg/2mosyqu.

It is reported that the AI will do in seconds what once took 360,000 hours of work performed by attorneys and loan officers each year. JPMorgan reports that COIN has helped the bank to minimize loan servicing mistakes, many of which are attributable to human error.

Law blogger Robert Ambrogi recently reviewed Joanna Goodman's book, Robots in Law: How Artificial Intelligence is Transforming Legal Services, for Above the Law. See http://bit.ly/2mgHdTz. Her book notes that the biggest market for legal AI is merger and acquisitions due diligence, in part because it is a sufficiently defined, narrow domain. She also points out that the biggest hurdle for AI in law is overcoming fear. "Ultimately, the incursion of AI into legal will relate to how far we are prepared to trust a 'robot lawyer.'"

Poor points to a McKinsey Global Institute study that predicts automation will create a "barbell" effect where it mostly benefits very large and very small companies by increasing agility for the former and scaling for the latter. See http://bit.ly/2jdlk70. Just like practice management software before them, it appears that robots are here to stay. Or, at bare minimum, they'll be back.


Matthew Hector
Matthew Hector is a senior associate at Woerthwein & Miller.