Legal analytics systems are tackling a vast and largely untapped ocean of data in Canadian legal decisions.
A number of Canadian startups are mining this data and giving it structure, converting what have otherwise been arcane systems into accessible statistics.
Toronto-based Loom Analytics is one such company, and it is using a combination of legal analysis and a type of artificial intelligence called machine learning to provide statistical analysis of Canadian case law.
The company has set about combing through every decision dating back to 2010 and plans to analyze cases before that once it has completed its work on court decisions in every Canadian jurisdiction.
Its reports include statistics on how specific judges have ruled on particular types of cases, the average length of time it takes them to arrive at a decision, as well as the frequency that specific types of claims make it to court.
Loom Analytics’ co-founder and president, Mona Datt, says providing structure to the data in Canadian case law allows people to make legal decisions based on math instead of gut instinct.
“All this data exists in case law and people aren’t really using it to make decisions consciously,” she says.
“Yes, they’re subconsciously bucketing things and doing a very long-winded process of actually doing case law research, but they’re not doing it mathematically.
They’re not able to tell a client ‘Well, there is a 70-per-cent probability you will do this.’”
With legal analytics systems, Datt says lawyers can actually tell clients how often a plaintiff has won in a specific type of case or how prone a judge is to ruling in a certain way.
“There is no ambiguity. There is no gut. There is no experience. It’s math,” she says.
Datt says her background as an engineer has given her a step up in being able to analyze the data as she sees things mathematically, whereas lawyers are more likely to see grey areas.
“If a lawyer came in and approached this, there would always be grey areas,” she says.
“Coming at it as an engineer allowed me to see the patterns, allowed me to actually see the differences. I was not approaching it as a legal problem. I was approaching it as a data problem,” she adds.
Legal consultant Jordan Furlong says that while giving more structure to legal data has a lot of benefit for lawyers, a potential drawback is that they might interpret the data within the “fairly narrow confines of the way we think as lawyers.”
He says bringing in people from outside of law, such as Datt, to analyze the data is vital.
“The solutions we come up with, there’s a risk that they will tend to perpetuate our own positions, advantages and practices,” he says.
“Bring in people from outside and they can see a little more clearly. We’re a little too close to the target. We have a vested interest that other people may not have.”
Tools that are providing structure to this data can save a lot of time and can be highly advantageous to lawyers from a research perspective, observers say.
Canadian startup Rangefindr provides criminal-sentencing ranges for specific fact scenarios based on a database of thousands of past decisions.
Lawyers who do this work manually often spend hours to create a detailed composite of sentence ranges for cases and they have to go back to square one every time as fact scenarios are often very unique, says Colin Lachance, vice president of development for Rangefindr.
“They don’t improve their process each time they go back and conduct more research,” he says.
“It’s like building a sandcastle that gets washed away or building a book that they have to tear up when they’re done because they can’t go back to the prior research.”
In addition to research, legal tech startups are finding other uses for the data that exists in case law.
Premonition is a Florida-based company that is using machine learning to rate lawyers based on a number of factors, including outcomes and the duration of trials.
The startup is planning to launch in Canada at the end of September and has already gone through approximately 65 per cent of Canadian court decisions.
Guy Kurlandski, the co-founder and CEO of Premonition, says that tackling all of that data has been challenging at times.
“There are a lot of structural issues in that data. It needs a lot of normalizing from territory to territory,” he says.
“We keep hitting what some people would probably consider to be brick walls because the task is so big and then we try to find ways to teach our machine to overcome them. Sometimes, the answers are simpler than they first seem.”
Fernando Garcia, general counsel for Nissan Canada, who is a member of Premonition’s board, says the development of artificial intelligence has been a huge part of developing this analysis as the sheer number of cases there are has made it very hard in the past.
“It’s one of those things where the advances in technology have made it feasible,” he says.
In addition to the absence of tools to analyze data easily, Furlong says a lack of access to case law has also served as a barrier to bringing more structure to this data.
It hasn’t been until CanLII started that widespread access to court decisions has actually been a reality, he says.
“Access is a fairly new thing and it’s not yet universal,” he says.
A lack of awareness or sensibility in the legal community to start questioning assumptions of how the law is typically done has also hindered the process, Furlong says.
“Self-awareness has not been one of the legal profession’s strongest characteristics,” he says.
A number of Canadian startups are mining this data and giving it structure, converting what have otherwise been arcane systems into accessible statistics.Toronto-based Loom Analytics is one such company, and it is using a combination of legal analysis and a type of artificial intelligence called machine learning to provide statistical analysis of Canadian case law.
The company has set about combing through every decision dating back to 2010 and plans to analyze cases before that once it has completed its work on court decisions in every Canadian jurisdiction.
Its reports include statistics on how specific judges have ruled on particular types of cases, the average length of time it takes them to arrive at a decision, as well as the frequency that specific types of claims make it to court.
Loom Analytics’ co-founder and president, Mona Datt, says providing structure to the data in Canadian case law allows people to make legal decisions based on math instead of gut instinct.“All this data exists in case law and people aren’t really using it to make decisions consciously,” she says.
“Yes, they’re subconsciously bucketing things and doing a very long-winded process of actually doing case law research, but they’re not doing it mathematically.
They’re not able to tell a client ‘Well, there is a 70-per-cent probability you will do this.’”
With legal analytics systems, Datt says lawyers can actually tell clients how often a plaintiff has won in a specific type of case or how prone a judge is to ruling in a certain way.
“There is no ambiguity. There is no gut. There is no experience. It’s math,” she says.
Datt says her background as an engineer has given her a step up in being able to analyze the data as she sees things mathematically, whereas lawyers are more likely to see grey areas.
“If a lawyer came in and approached this, there would always be grey areas,” she says.
“Coming at it as an engineer allowed me to see the patterns, allowed me to actually see the differences. I was not approaching it as a legal problem. I was approaching it as a data problem,” she adds.
Legal consultant Jordan Furlong says that while giving more structure to legal data has a lot of benefit for lawyers, a potential drawback is that they might interpret the data within the “fairly narrow confines of the way we think as lawyers.”
He says bringing in people from outside of law, such as Datt, to analyze the data is vital.
“The solutions we come up with, there’s a risk that they will tend to perpetuate our own positions, advantages and practices,” he says.
“Bring in people from outside and they can see a little more clearly. We’re a little too close to the target. We have a vested interest that other people may not have.”
Tools that are providing structure to this data can save a lot of time and can be highly advantageous to lawyers from a research perspective, observers say.
Canadian startup Rangefindr provides criminal-sentencing ranges for specific fact scenarios based on a database of thousands of past decisions.
Lawyers who do this work manually often spend hours to create a detailed composite of sentence ranges for cases and they have to go back to square one every time as fact scenarios are often very unique, says Colin Lachance, vice president of development for Rangefindr.
“They don’t improve their process each time they go back and conduct more research,” he says.
“It’s like building a sandcastle that gets washed away or building a book that they have to tear up when they’re done because they can’t go back to the prior research.”
In addition to research, legal tech startups are finding other uses for the data that exists in case law.
Premonition is a Florida-based company that is using machine learning to rate lawyers based on a number of factors, including outcomes and the duration of trials.
The startup is planning to launch in Canada at the end of September and has already gone through approximately 65 per cent of Canadian court decisions.
Guy Kurlandski, the co-founder and CEO of Premonition, says that tackling all of that data has been challenging at times.
“There are a lot of structural issues in that data. It needs a lot of normalizing from territory to territory,” he says.
“We keep hitting what some people would probably consider to be brick walls because the task is so big and then we try to find ways to teach our machine to overcome them. Sometimes, the answers are simpler than they first seem.”
Fernando Garcia, general counsel for Nissan Canada, who is a member of Premonition’s board, says the development of artificial intelligence has been a huge part of developing this analysis as the sheer number of cases there are has made it very hard in the past.
“It’s one of those things where the advances in technology have made it feasible,” he says.
In addition to the absence of tools to analyze data easily, Furlong says a lack of access to case law has also served as a barrier to bringing more structure to this data.
It hasn’t been until CanLII started that widespread access to court decisions has actually been a reality, he says.
“Access is a fairly new thing and it’s not yet universal,” he says.
A lack of awareness or sensibility in the legal community to start questioning assumptions of how the law is typically done has also hindered the process, Furlong says.
“Self-awareness has not been one of the legal profession’s strongest characteristics,” he says.
