Over the past year, legal technologists concerned with access to justice have increasingly focused on the importance of the interface in bringing legal resources to under-resourced communities. Chatbots like DoNotPay leverage instant message like chat to help people access legal content, while others like CuroLegal have tried to beef up simple guided interview interfaces to provide information.

Legal network service provider Legal.io is concerned with these same interface questions for its new legal assistant tool, LARA. Using machine learning, the tool interprets both voice and text language an average person might use to ask a legal question, then suggests relevant data from Legal.io's resources.

Pieter Gunst, chief operating officer at Legal.io, said the company took note of the growth of artificial intelligence (AI) infrastructure in legal enterprise technology and looked for an opportunity to apply it to the company's consumer base.

“We're more focused on what these tools can actually do for consumers in terms of navigating the legal system,” he said.

At present, LARA has a couple core functionalities. It primarily produces non-disclosure agreements (NDAs) by taking relevant information in by voice or text. Its other functions—finding an attorney and pro bono services—are still “mostly experimental,” according to LARA's website. Gunst noted that Legal.io is currently working on ways to delimit the kinds of information LARA can draw from so that the company's bar association partners can use the tool in more regional or practice-specific ways.

NDAs are a popular starting point for organizations looking to build out broad document generators. LISA, another female-named document preparation tool designed by Chrissie Lightfoot, president of EntrepreneurLawyer, and Adam Duthie, partner at Duthie & Co., also used NDAs as a foothold into a broader development of document creation tools.

Although technologists tend to tout their use of one or two core technologies, Gunst noted that effective strategies tend to pull from lots of different technology structures. “What we've found is to get a useful response, you're kind of using a combination of technologies. You're using search, you're using a part of a document generator,” he noted.

LARA currently can take inquiries via chat, Facebook Messenger and voice, but Gunst noted that the Legal.io team is paying close attention to the ways in which LARA users feel most comfortable accessing their information and experimenting with some new interfaces.

“You're likely to see some of that in LARA later. We're experimenting with what is the preferred way for a consumer to use this,” he said.

While the Legal.io team is working to create interface access points for users, as with any of these tools, machine comprehension isn't always totally there. My LARA inquiries for a local immigration attorney sometimes put me in the wrong state, or occasionally produced no attorneys at all. But, also like any of these tools, the machine learning insight of the tool can sometimes take some time to build.

Legal.io has that long-game in mind, with Gunst saying, “In our minds, it's an experiment around what's possible.”

Over the past year, legal technologists concerned with access to justice have increasingly focused on the importance of the interface in bringing legal resources to under-resourced communities. Chatbots like DoNotPay leverage instant message like chat to help people access legal content, while others like CuroLegal have tried to beef up simple guided interview interfaces to provide information.

Legal network service provider Legal.io is concerned with these same interface questions for its new legal assistant tool, LARA. Using machine learning, the tool interprets both voice and text language an average person might use to ask a legal question, then suggests relevant data from Legal.io's resources.

Pieter Gunst, chief operating officer at Legal.io, said the company took note of the growth of artificial intelligence (AI) infrastructure in legal enterprise technology and looked for an opportunity to apply it to the company's consumer base.

“We're more focused on what these tools can actually do for consumers in terms of navigating the legal system,” he said.

At present, LARA has a couple core functionalities. It primarily produces non-disclosure agreements (NDAs) by taking relevant information in by voice or text. Its other functions—finding an attorney and pro bono services—are still “mostly experimental,” according to LARA's website. Gunst noted that Legal.io is currently working on ways to delimit the kinds of information LARA can draw from so that the company's bar association partners can use the tool in more regional or practice-specific ways.

NDAs are a popular starting point for organizations looking to build out broad document generators. LISA, another female-named document preparation tool designed by Chrissie Lightfoot, president of EntrepreneurLawyer, and Adam Duthie, partner at Duthie & Co., also used NDAs as a foothold into a broader development of document creation tools.

Although technologists tend to tout their use of one or two core technologies, Gunst noted that effective strategies tend to pull from lots of different technology structures. “What we've found is to get a useful response, you're kind of using a combination of technologies. You're using search, you're using a part of a document generator,” he noted.

LARA currently can take inquiries via chat, Facebook Messenger and voice, but Gunst noted that the Legal.io team is paying close attention to the ways in which LARA users feel most comfortable accessing their information and experimenting with some new interfaces.

“You're likely to see some of that in LARA later. We're experimenting with what is the preferred way for a consumer to use this,” he said.

While the Legal.io team is working to create interface access points for users, as with any of these tools, machine comprehension isn't always totally there. My LARA inquiries for a local immigration attorney sometimes put me in the wrong state, or occasionally produced no attorneys at all. But, also like any of these tools, the machine learning insight of the tool can sometimes take some time to build.

Legal.io has that long-game in mind, with Gunst saying, “In our minds, it's an experiment around what's possible.”