Guest post: The hidden clause – The poisoned needle in a haystack

By Hugo Seymour     

A lot has been said about how today’s in-house legal departments can improve and modernise by becoming more data-driven and, of course, more agile. Much of this commentary has been made in the context of the increasing pressure that business leadership and commercial teams are exerting on their legal teams and how technology will change things. So, how will technology, specifically the latest iterations of generative AI help?     

Reviewing contracts has been an early and successful use case for AI in legal. The use of supervised and unsupervised machine learning to review and extract insights from multiple legal documents has reduced the amount of time and (human) resources it requires to complete that process. While we need to acknowledge that in-house teams do a lot more than just review contracts and legal documents, how will today’s AI help them do that part of their role even more efficiently and allow them to focus on more value-add activity?      

Finding a needle from the haystack     

With contract review the cartoon, nightmare scenario is that you, the board, and many employees go to prison. However, the reality is that familiar nagging concern that there is something ‘you do not know’ loitering inside your business’s contract stack, waiting to cause trouble. Not just a ‘needle in a haystack,’ but a poisoned needle that needs to be found before it causes trouble. No one likes nasty surprises, no more so than a legal professional whose job it is to review, analyse, and weigh up the risks and to avoid being hit by a litigation landslide.     

Unfortunately, the unforeseen happens more often than we might expect, and ‘bad’ terms or hidden clauses often leave businesses in bad commercial positions. However, hidden clauses are an interesting concept because no clause is ever truly hidden. It is just not easy to find. This is especially true when you consider the way most legal departments review contracts, which typically consist of legal teams manually reviewing hundreds of contracts. As contracts become increasingly complex and numerous, it is more difficult to discover all the ‘hidden’ information within a contract manually.      

Is there another way?      

One way to mitigate the challenge of finding hard to find, nasty surprises in your contracts is to use artificially intelligent question-answering (QA) technology, which enables legal professionals to analyse documents based on their own questions and checklists, which allow them to make the best use of their own expertise to find any troublesome needles. So, rather than searching for the existence (or non-existence) of specific clauses to determine the right course of action, you can ask your contract management system a simple question (like you would do in Google’s search engine) and get a clear answer that has been pulled from your contract base.     

Similarly, if you ask a question and no answer/solution is available, it clearly indicates that this crucial information is missing. Knowing that could save you many wasted hours searching for the elusive ‘needles’ and allows your searches to reflect the nuance and specificity of real-life needs.     

Tidying up language ahead of any negotiations is also much easier. We all know starting a discussion about a specific point in contact during negotiations is not wise, so knowing that you might have a problem with a particular clause in advance allows you to deal with it in a way that does not put you at an immediate disadvantage.   

Extracting the needle from the haystack     

The ability for people to be clear about what they are looking for, and the increasing power of machines to find it, will make a massive difference for in-house counsel. They can find and deal with risks and practical problems at their own pace and, crucially, at their chosen time. The benefit of the move of these tools in-house is that they will not need to get external counsel involved.   

Using these tools in the right way starts to create data. Specifically, it starts to generate some actionable data, but we will talk about why that is important in the following article.   

Hugo Seymour is AI/ML product lead at Della, which was acquired by Wolters Kluwer in January 2023.