We’re long past the early days of NLP in the first search engine and spell-check. Now an essential component of contract management, NLP is going to have an increasing influence on the way you do legal.
Natural Language Processing (NLP) has been part of your life since the early 1990s. Spell check and simple searches (remember Ask Jeeves?) was one of the first applications, now it’s omnipresent. Email, Alexa, Siri, chatbots, search results, translations – almost everything you do online has some form of NLP involved.
NLP is a facet of Artificial Intelligence that helps computers understand, decipher, translate and extract meaning from human language. Learning words is easy for computers; it’s understanding meaning, context, and sentiment where the challenges occur.
When we speak human-to-human there are multiple factors beyond simply words – intonation, sentiment, volume, body language, to name just a few. That’s why often that seemingly snarky email from Claire can be so easily misinterpreted.
When it comes to legal – a sector that is founded on a love of language and where the most minute ambiguity can mean the difference between winning and losing a case – NLP has never been so prevalent or so important. A recent report by research firm MarketsandMarkets predicts that the NLP market will reach $2.64 billion by 2024.
But what’s driving this market growth and what’s next for NLP and legal? Here are five NLP trends worth watching:
Chatbots: back in 2011 Gartner predicted that by this year, 85% of customer service would be done without human interaction. The jury is still out on whether that prediction has come true. But one thing we can sure of is the rise of chatbots – those annoying boxes that pop up when you go on to your bank, or a sales-led website asking if you’d like to ‘chat now.’ The Law Society is already running articles about how chatbots can increase your business, either by helping those people who can’t speak directly to a lawyer for minor issues, for legal research purposes, or to help streamline client-enquires and filter them to the right attorney.
Answers, not results: no doubt you’re already using a contract management system that involves some level of NLP (and if not, why not?). But when we search for terms or clauses, the NLP unleashes results, not answers. It delivers a stream of information – much like Google before its Google Quick Answers was developed – which then requires human intelligence to decipher and sort through to discover the answer to the question. NLP is developing at such a rate that it will be delivering answers to our questions, not just all the information that it thinks might be relevant. From a legal perspective, obviously, this has pros and cons for areas like E-Discovery, where less information isn’t always an advantage, but from an efficiency perspective, it could be a huge business advantage.
Company sentiment analysis. due diligence is a laborious, often expensive and always a tedious necessary evil. One of the challenges is that the business landscape is a constantly moving beast; what’s part of your CSR program one day is greenwashing the next. Getting this level of detail in real-time during the process is challenging and yet could impact any possible sale. NLP will become more and more essential for law firms conducting due diligence and accessing market sentiment through news, social, and markets that could have a potentially huge impact on share or sales prices in real-time. This means GCs and law firms are fully prepped to give up-to-the-minute advice to impact strategic decisions.
More complex questions solved: as HBR says, answering one simple question – ‘Who are my customers?’ – is relatively easy for NLP. But answering more complex, layered questions – ‘Who is my best customer in the northwest for contract management’ – is an altogether more challenging prospect. The algorithm needs to understand a definition of ‘best,’ have a concept of what region constitutes the North West, and know what products make up a contract management offering. According to HBR, just adding one facet to the query increases the complexity exponentially. Over the next several years this will change dramatically. Data scientists already have multitask question-answering models, which means they can tackle layered questions and complete tasks that haven’t been seen before or are not trained to do.
It’s all about Sesame Street: without going deep tech, there have been a couple of advances in NLP over the past two years that will have a long-lasting impact. BERT (Biodirectional Encoder Representations from Transformers) contextualizes words – i.e. it checks the words around a given word to ensure the correct understanding. ERNIE (Enhanced Representation through KNowledge IntEgration), combines BERT with other outside information, such as encyclopaedias, news sites, and online forums. This means the NLP also uses other external information to make better sense of language. For legal, this means a leap in accuracy in understanding terms, clauses and conditions. It also means greater efficiencies – faster time to read documents, more accurately with less human intervention – and therefore the ability to offer clients in-depth strategic legal counsel, not just document review.
The jury isn’t out on how important NLP is going to be to your organization. It’s back and it has given the final verdict. NLP is an integral part of our lives and the law. Given the progression we’ve seen in the past two years alone, this symbiotic relationship is only going to become deeper, changing the way legal operates.
For information on how Exigent maximizes the benefits of NLP within its own Machine Learning and Contract Management Solution, click here.