AI-Powered Legal Assistant Finds ‘Trusted’ Lawyers, Answers Legal Questions

A new virtual legal assistant, LARA, uses artificial intelligence and millions of data points to help consumers find legal help and get answers to legal questions.

Developed by Legal.io, a company that builds attorney referral networks for legal organizations, the purpose of LARA is to use data to help clients find trusted legal help.

With this alpha-stage launch, LARA’s “skills” are primarily limited to the following:

  • Get a non-disclosure agreement.
  • Find a lawyer referral service.
  • Find a legal aid organization.
  • Find a lawyer

LARA can also answer several thousand legal questions, mainly derived from public sources, such as the courts. Legal.io CEO Pieter Gunst says that the company will be monitoring user queries, and will be adding relevant skills over time.

But the larger goal, as Gunst explains in this Medium post, is to tap into data to improve the process by which people find a lawyer. LARA is not just a robot that answers legal questions, but a service that helps them find the right legal help “to fully handle their most important, sensitive needs.”

Because despite technological advances, the process of finding the right legal help remains daunting and inefficient. It’s hard to find out which attorney is good at what. It’s near impossible to find accurate pricing information. Most people perceive the cost of legal help to be significant at best, prohibitive at worst. In short, that data is rather hard to come by, and as a result, the legal marketplace isn’t very efficient.

Through its work with legal associations, Legal.io has built an attorney network that rivals the size of the largest law firms, Gunst says, and that has facilitated millions of dollars worth of referrals.

LARA draws from this, using AI to pull from data sources and applications powered by the Legal.io platform.

Ask LARA about an NDA, and it will help you create one, if you want, using a bot to walk you through the necessary information. But it will also help you find an attorney in your area who can assist with creating an NDA at a flat rate, or help direct you to a lawyer referral service in your area. It also gives you an overview of fees charges for NDAs, showing both the range and the median.

Not of this is particularly novel, Gunst concedes. Other services help people create forms and find lawyers. But he says the vision is larger, in part because it is rendering results that are updated in real time.

Instead of a static price posted on a landing page, pricing information is pulled directly from hundreds of attorneys across the globe. Instead of Avvo telling attorneys what to charge for an NDA, our platform collects thousands of legal data points and collates that information in a single location.

Further, it’s more than just another chatbot or document generator. It gives the client a direct way to connect with a trusted attorney or local bar association lawyer referral service.

Gunst emphasizes that LARA is still experimental. But the idea is a good one. A frequent criticism of legal bots is that they do not and cannot provide sophisticated legal help. LARA is different, in that it offers consumers self help but also the ability to easily find actual legal help and to have an understanding of what it will cost them.

Posted in:
Updated:
  • pietergunst

    Thank you for writing about this!

    Would love feedback and thoughts from the community. We’re taking an agile approach, and also want to be clear that the capabilities of LARA (or any AI system today, really) are rather limited. I’m noting this, because there has been so much hype around AI and bots, and part of our aim is to share learnings, and to engage in open conversation about what is and what is not possible today, and what approaches really do benefit the consumer of legal services.

    The combination of AI and micro-applications (whether rendering relevant attorneys, pricing info or a self-help tool, such as the NDA bot), is I believe particularly promising in helping clients navigate the legal system. That being said, we expect that a human-handoff (connecting the client with an attorney, or where necessary, a legal aid resource) will still be the best solution in the vast majority of use cases. At the same time, effective triage can save time both on the end of the consumer, and the legal professional.

    For the more tech-inclined readers: the tech stack for LARA is in house AI (although we’ve also used TensorFlow and Microsoft LUIS). We’re using Bot Framework for the NDA bot (Node.js). Data is pulled from various segregated databases provided by the Legal.io platform, which also provides the general infrastructure. LARA’s been developed as a JS widget, so it can eventually be dropped into any website, whilst pulling information from any Legal.io implementation.

    Also happy to receive private questions / feedback, via pieter@legal.io, and thank you for reading!

    • Bob Ambrogi

      Thanks for adding this additional information.