fivehundred magazine > Technology > How to become a smart shopper of AI

How to become a smart shopper of AI

Whether making a direct investment or utilising the products of others, there’s plenty of options for robot lawyering, says Dan Carmel, chief marketing officer of iManage

illustration of robot shopping

Information is the coin of the realm for any corporate legal department or law firm – and that is probably why the topic of artificial intelligence (AI) has a generated so much excitement across the legal industry.

AI, properly applied, can produce dramatic productivity improvements, enabling document reviews and other basic legal tasks to be done much faster, with greater accuracy and at lower cost, than people. But, the improper application of AI can result in project delays, work that has to be redone manually, alongside disappointed and disillusioned lawyers. How do you avoid these pitfalls and realise the benefits of AI? By being a smart shopper.

As a smart shopper, let’s start by understanding where AI ‘fits’ in legal and where it does not. AI takes many forms, and so I will focus here on machine learning and cognitive task automation – the two most frequently cited areas of AI applications in legal. In general, the more specific the task or project to be automated, the better the result from AI. For example, finding parties named in all asset purchase agreements, versus finding the parties in any kind of contract.

Also, the more similar and repeatable your documents are, the better the results. So, if the leases you want to analyse have similar form, the results you get will probably be of higher quality than if they vary dramatically. Therefore, AI is best applied to consistent document sets, with high volumes, and where you can expect good commonality.

Second, AI is not going to replace lawyers. It will augment lawyers’ work by automating tasks that are unpleasant, like detail document reviews. We describe the use of AI today, as a man-machine partnership, not too dissimilar from how you use a dishwasher or a washing machine. These machines don’t do the whole job, but they take the drudge work out of the bigger job. That is the state of most legal AI in 2019. Good judgement is not a capability of AI yet.

Another consideration is whether you want your organisation to be expert in applying the AI technology, or outsource it. If you’re thinking of developing a competency ‘in-house’ for AI, be prepared to invest in recruiting and training the right staff to apply this technology to different matters or practice areas. You’ll need to think carefully about skill sets, staffing, and ensuring the technology you use can be easily maintained and integrated with other systems.

If you are not prepared to undertake this kind of investment, because your volumes of documents are not large enough, or, like many legal departments, you don’t want to hire the dedicated expertise for such a specialised area, you will end up encountering AI in a less-direct way, such as:

  • Licensing AI-based applications and legal services from law firms. More and more law firms are beginning to develop repeatable products and services based on a combination of AI, specific target document types, and custom legal logic. These offerings are made available to legal departments, and sometimes to other law firms, and are intended to be close to a ‘turnkey’ solution.
  • Encountering AI as feature of a software application you already use. A lot of software applications in areas such as deal management, document management, collaboration, and search are being augmented by AI technologies to make them smarter and increase their value. In these instances, the AI is not exposed to you – it is embedded inside another software product and merely automates the tasks for which it is intended. We see this as a dominant application of AI in legal over the next five years, as it is task specific, repeatable and delivers high value, without the cost of a bespoke solution.

For example, smart document management provides a terrific example: embedded AI can classify documents by document types; mine documents for clauses and recommend best drafting language; extract key terms and entities from documents, which enables better analytics; search and expertise identification, for more effective content re-use; and knowledge management.

AI-powered security products provide another example: they can use machine learning to build a digital fingerprint of individuals to understand their individual behaviour profiles of professionals and then issue an alert when that behaviour deviates from expected patterns – helping identify and stop data breaches and insider threats much more quickly, than traditional security approaches.

Whether making direct investment in AI solutions, purchasing services from AI-enabled law firms, or using smart products that have embedded AI, there’s an ideally suited way for law firms and legal departments to start boosting their AI quotient based on their individual requirements. All it will take is a little smart shopping.