Understanding the complexity of automated contract review

Contracts are the heart of most transactional workflows. But for many firms, not much has changed over the years in how documents are reviewed. The process is still overly manual. As a result, lawyers spend a significant amount of time reading and extracting data to create and negotiate contracts of all levels of complexity to ensure their clients are not exposed to risk.  

To make contracting quicker, more accurate, and more predictable, firms are now starting to embrace contract automation. In fact, Gartner predicts that by 2024, advanced contract analytics solutions will reduce the manual effort required for contract reviews by 50%. Machine learning will play an important role in this evolution, as it will emulate human decision makers when processing contracts. But these machines must first be programmed to think like a lawyer. That means thoroughly understanding what’s involved in the contract review process.

“The concept of reviewing a contract is broad and can look very different from transaction to transaction,” says Steve Fullerton, Product Manager at Thomson Reuters. “But we also find lots of common pain points.” These shared attributes are where we need to focus.

The nuances of contract review 

Before tackling the range of use cases suitable for technology-assisted contract analysis capabilities, let’s look at the tasks associated with various stages of contract review:

• Not all information is simple to find. The answers to questions can be in several locations within a document or across related documents. It’s not always where one might expect. This search becomes even more complex when scaling document reviews. Clauses in different locations, even different documents, can alter or supersede initial meanings. To pinpoint the actual meaning, lawyers must recognise the logical relationship between multiple references.
• Language and legal meanings can vary. Lawyers look beyond the text and compare meanings, not just language. The interpretation of legal meaning, in particular, is not always cut and dried Some lawyers may determine that meanings expressed differently in words are effectively the same. Similarly, two lawyers can disagree on a meaning or interpretation. If this is a challenge for humans, imagine how much harder it would be for a machine to learn differences such as these.
• Context awareness will impact risk. Reporting, rather than review, is the ultimate end goal. “The process of any type of review is to understand what these contracts mean,” says Fullerton. “Once you’ve learned what a contract means, you then need to interpret the meaning for your client. The context of every transaction will impact risk, so the report will need to translate the legal language into business language to clarify risks and materials and match policy positions to the client’s playbook.”

Watch our HighQ Contract Analysis launch webinar here to see these tools in action.

An opportunity for machine learning

Lawyers spend a lot of time on both transactional and post-transactional issues. Where there is a task that consumes a large part of their workweek, there’s an opportunity for automation to ease the burden.

Artificial intelligence and machine learning can help lawyers tackle new and existing challenges around risks, workflows, and reporting. They can easily extract data and clarify the content of contracts, allowing firms to review documents more rapidly, decrease the potential for contract disputes, and increase the volume of contracts they can negotiate and execute. Once the challenges are understood, the next step is to identify the right techniques to apply to deliver value.

Our next blog will discuss what these machine learning techniques are and where they can be applied to simplify the contract review process.