AI takes hold in the legal profession

Professional workflows

Automating workflows in professional services in general has been challenging because the processes are often not simple or repetitive; they require an intelligent workflow and usually intervention by a human at some point. Autto was developed to automate portions of legal work, up to the point where human intervention is required, and is being modified continuously to provide new automated processes.

“Bright, well educated (and expensive) lawyers are spending 20 to 50 percent of their time doing things that do not take advantage of or require their judgment or education,” says Ian Gosling, CEO of Autto. “This contributes to the very low job satisfaction ratings even from senior lawyers and also reduces efficiency.”

Lawyers have accepted workflow for case management, but not for the conduct of their tasks. “There is a cultural barrier,” Gosling says. “Lawyers are trained almost like artisans to craft a unique solution for their client after listening carefully to their situation. The idea of automating any process has been difficult to accept.”

Autto’s team consists of a lawyer, an entrepreneur and a software developer, all of whom met at another startup and decided to launch their own. Gosling explains, “We are getting traction because we focus on micro-automation—we are not trying to automate a full process, but to get the process to a point where it needs a lawyer’s critical faculties.”

He continues, “If you look at what lawyers are doing, there are some common building blocks. At the most abstract level, they are gathering information, putting conditions on it and then there is a resulting action. Going down another level, you can gather information from a variety of sources, including a person, a system or from an analysis of documents. The resulting action may be an approval, sending an email or putting information into a system.”

Autto allows lawyers to map their own workflow, gather information in a set of interactive forms, set conditions and trigger actions or if necessary, more information gathering. Workflows Autto has developed for clients include:

• providing guidance on immigration options based on top-level client information,
• creating the first draft of investment agreements for startup companies, and
• generating guidance and drafting paperwork for compliance with employment law.

Gosling believes that the next generation of lawyers, which he refers to as Law 2.0, are disillusioned with the existing model of big law firms. “They want to create efficiencies for their clients and provide a better service. It is much easier these days to set up a small legal practice because of the technology tools and the availability of subscription-based software,” he says.

Co-founder Max Cole cites figures indicating that the number of legal technology startups in the United States rose from 15 in 2009 to 400 in 2016. Uptake of technology has shown a significant increase, particularly in the last year, and may be correlated with the alternative business structure (ABS) now available in England and Australia, which allows non-lawyers to own or invest in law firms. “The shift over the past four years has been like night and day for legal technology,” says Gosling.

Preventing misdirected email

Most of us know the dreaded feeling of having clicked “Send” and then realizing the email has gone to the wrong person. For a lawyer, the feeling is more than one of embarrassment; it is a professional hazard with potentially severe consequences, including penalties for data loss. CheckRecipient was founded to address this problem by verifying that the email is going to the right individual.

“Organizations in the U.K. have to report personal data breaches to the Information Commissioner’s Office, which publishes statistics on these breaches every quarter,” says Tim Sadler, CEO and co-founder of CheckRecipient. “What surprises a lot of people is that misdirected emails are the number one digital data security incident reported.”

Based in London and founded by ex-investment bankers, engineers and mathematicians, CheckRecipient was developed as a result of Sadler’s observations about inadvertent data loss during M&A projects. “Email is the main channel of communication and is highly prone to human error,” Sadler says. “We apply machine learning to properties of email, and our software determines whether the message is appropriate for the recipient in real time.

The software works in the background and sends an alert to the user if any anomalies are detected in the outgoing email. The user’s options are determined by how the software has been set up by the administrator. “For many companies, our software is the first manifestation of AI in their organization,” Sadler says. Focused on one thing, CheckRecipient addresses a high priority need in legal offices.

Although CheckRecipient may be the first foray a law office makes into AI, it’s unlikely to be the last. AI applications for the legal profession are up and running in many offices. Corporate legal departments are bringing work in-house that was previously done by outside firms to contain costs, and they have an incentive to perform their work more efficiently. Despite the hourly billing that still is the norm in U.S. law offices, they too may become more receptive to technology as they feel continued pressures on costs.

It is probably no coincidence that so many AI products for legal use have originated in the United Kingdom. The ABS model has brought a more innovative mindset to the practice of law, and the push is on for re-invention. As for the impact on jobs, the outlook remains uncertain. Lawyers will not be replaced by software products anytime soon, but if the time to carry out certain tasks is dramatically reduced, change should be expected. Lawyers may be able to devote more time to higher-level activities or expand to underserved groups. In any case, legal AI is a space to watch.