The business world has a problem, and it’s one most people outside the office wouldn’t guess. The problem is that we’re drowning in contracts and don’t have enough people to read them. Even reading isn’t the whole problem. Rather, it’s more like “parsing through twenty contracts looking for one clause that could invalidate the other nineteen.” This isn’t an unusual scenario. It could happen hundreds of times per day.

Contract management has been an issue for businesses of every size, for at least the last few decades. In the business world, most companies have a Procurement department, used to obtain goods, services, or works from exterior sources based upon agreed terms. Contract management for Procurement alone becomes a daunting task with even thirty contracts on file. Out of necessity, Procurement departments everywhere have turned to machine learning AI, which goes a step beyond merely filing contracts. It reads them.


What is machine learning?

The phrase “machine learning” still sounds odd, even though we’ve been using that term for a few years now. We know that machines can do work, and we know that they have to be programmed to do that work. Isn’t “machine learning” just teaching it a new program?

Not exactly. To write a computer program, we must first understand how we do an action and then break it down into steps for the computer to perform. But we don’t know how to tell a computer to, for example, recognize a face, even though a baby knows how to do it. It’s something we do without verbalizing the steps. Machine learning, instead, would be a process of flashing thousands of face images at the computer and telling it which ones are the correct person, until the computer knows how to recognize that person from any camera angle. This is only possible because computers are fast enough now to process so much raw, unfiltered data.

In the same way, computers can be taught to “read,” or at least to sort through pages of text and apply learned language pattern algorithms to identify what parts of a contract are a date, a dollar amount, a named party, a forfeiture condition, and an exception clause. It can then deposit that sorted and tagged data until a human can ask it “What terms of this contract don’t agree with the other nineteen?” Then it puts up a handy graph that answers the question.

How can machine learning help contract management?

Contracts for any company have become an awkward workload that nobody seems to want around. For instance, in this survey, 35% of legal professionals say they think contract management is a legal responsibility, 45% that it’s a business responsibility, and 20% without even that much conviction for an answer. Meanwhile, managing these contracts amounts to a maintenance hassle of which 65% of legal professionals report being fed up.

An AI that helps plow through this workload is an incredible help to businesses of every size. Companies that have taken up the first wave of machine learning adoption have seen their business perform at top shape, leaving the competition behind. This makes sense when you realize that an AI can now beat lawyers in a competition to find errors in a contract. Human error in conducting contractually bound business obligations cost billions in losses annually. Even if a company is assured of their contract’s terms, discovering when the other party has breached its obligations is a matte of intense review.

Like many other aspects of business, machine learning will one day be said to have changed the face of contracted companies forever.

 

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.