Company’s software analyzes and summarizes legalese into graphical charts.
Read in 2 minutes
Cian O’Sullivan has a simple way to test the effectiveness of the contracts he writes. He’ll put one specific clause into the document, then ask the reader if it’s in there or not. Most of the time, the reader can’t say one way or the other despite having gone through it.
The example highlights the problem with many modern contracts, including those that virtually every internet user agrees to when they sign up for an online service such as email or music streaming. Most are so long-winded and full of legalese that the average person has no idea what they’re actually agreeing to.
The obvious answer is simpler contracts, but with companies needing to protect themselves from every conceivable eventuality, it’s not that simple. The better solution, O’Sullivan believes, is artificial intelligence, which is why he started his company Beagle three years ago.
“People are really good at making complex, subjective decisions when all of the information is presented to them,” he says. “We’re trying to reduce the noise and bring out those most important bits.”
Beagle’s software uses natural language processing to analyze contract language. It then picks out the important parts and delivers them in an easy-to-understand graphical summary. The software can identify each party’s key responsibilities and liabilities, as well as clauses that can lead to termination:
The idea for Beagle came from O’Sullivan’s owns experience as a contract negotiator for Desire2Learn, an educational technology company based in Kitchener, Ont. He came to dislike the tediousness of constantly having to explain contracts and their terms to clients.
“It was effectively [like the movie] Groundhog Day. I was doing the same thing over and over again,” he says. “I had to believe there was a way to bring those salient points out of a contract and bring them to the top.”
He started writing software that could identify key phrases, but also important concepts, and eventually attracted others to his cause. They settled on Beagle as the name for their company, also based in Kitchener, because of the dog’s keen sense of smell. It matched the AI’s ability to sniff out important clauses.
The company, which now has 14 full-time employees, has been primarily self-funded so far but has also attracted investment from Next Law Labs, the venture development company belonging to legal giant Dentons.
O’Sullivan was originally planning to sell Beagle as a subscription service to small- and medium-businesses, but large enterprises have also caught wind. One of the largest auto makers world, which he says he can’t name for confidentiality reasons, is a customer.
Beagle’s particular flavour of AI is the non-threatening kind, he adds, in that it isn’t likely to jeopardize any job that a human wants to do. Flipping through contracts and understanding them is, after all, a task most businesses would prefer to fully automate.
“We’re using technology to do things that people are not necessarily good at,” O’Sullivan says. “You’re reducing error rate by using effective technology.”