Leveraging chatbots

The size of the chatbot market is growing exponentially. Gartner forecasts that, by 2020, more than 25 per cent of customer service operations will be handled without a human. There continues to be fewer human interactions in our daily life, and we will be interacting more and more with automated and life-like computer programs. Some of us already engage with several bots daily — from Siri to Google Home to Alexa — or when we call into customer service lines. These bots mimic human conversation interaction and, from the back end, operate almost like a choose-your-own-adventure book (maybe I’m dating myself here).

Online chatbots are not just client-facing but can be used to automate customer intake online and other business tasks. In the legal space, chatbots can be used to collect information to automate legal document creation and enhance client experiences. For lawyers, chatbots could be leveraged while delivering these services at a lower cost, for around-the-clock and just-in-time services.

What is a chatbot?

A chatbot is a computer program that is pre-programmed with certain rules (rules based on artificial intelligence) or trained to conduct a defined conversation (machine-learning artificial intelligence). The chatbot could be voice activated, like Alexa or Siri, or it could be text based like a weather bot. A Juniper Research report estimates that chatbot adoption in banking, retail and health care will be responsible for cost savings of more than $11 billion annually by 2023, up from $8 billion in 2017. In fact, the PwC report, “Bot.Me: A revolutionary partnership, How AI is pushing man and machine closer,” found that 31 per cent of business executives believe that more than any other AI-powered solution, virtual personal assistants will have the largest impact on their business.

Law and chatbots

To consider how chatbots could be used in law, I am going to take you back to last fall. On Sept. 7, 2017, Equifax reported that 143 million U.S. customers may have been victims of a cyberattack. This data breach saw the theft of millions of names, birthdates, social security numbers and addresses — all the information a cybercriminal would need to commit identity theft. In the aftermath of this breach, many people were left wondering about their legal rights. Enter DoNotPay.

DoNotPay is a legal technology that uses an artificial intelligence software (chatbots) to assist consumers with different (primarily small-dollar-amount) legal issues, such as fighting parking tickets or flight refunds. The software proceeds to have a conversation with the user to determine the best course of action for each specific situation.

In the days following the Equifax breach announcement, DoNotPay’s founder, Joshua Browder, announced a version of the DoNotPay app that would provide people affected by the data breach with legal information and completed PDF forms that could be used to file a case in court. The software provided a way for millions of Americans to understand their legal rights and prepare legal court documents for free. However, critics of the software said that it didn’t adequately advise or counsel people on how to quantify damages suffered, if suffered at all, and did not advise them that they would not be able to participate in a future class action lawsuit.

Criticism aside, DoNotPay has garnered quite a bit of attention in the legal technology space because of its use of AI and its focus on access to justice. Imagine all the small-dollar-amount legal issues that your family or friends come to you about, such as overcharges on a phone bill, being rebooked on a flight or a parking ticket. It doesn’t make economical sense for a lawyer to spend time on these issues, when they could be assisted cheaper by AI bots.

Another legal chatbot is Destin.ai, a Canadian-based legal software that describes itself as an “immigration assistant” in Canada. The welcoming and easy-to-understand website employs a Facebook Messenger chatbot to determine the eligibility of the inquirer of immigration to Canada.

Chatbots can be used not only to provide legal information to the public but also to automate and improve marketing or daily tasks done by a lawyer. One such software is LawDroid, which assists lawyers to
“[s]chedule appointments, dictate notes, create and assign tasks, all with the power of your voice.” As opposed to DoNotPay, LawDroid is lawyer facing and can assist lawyers to create their own chatbot for their websites. An around-the-clock chatbot could assist in converting website visitors to leads and standardize law firm tasks, among other things.

Create your own chatbot

The future is bots, and lawyers can take advantage of them to improve their law firm service. To create your own chatbot, you will want to:

Define your goals: What will your chatbot do? How will you measure success? What possible benefits in your practice would there be? How does that compare to the cost of creating the chatbot?

Choose a channel: Where do you want your chatbot to operate? Consider how your clients currently communicate with you and/or what platform they are more likely to use — do you have a Facebook page, do your clients use Facebook Messenger or is your website the more appropriate place to engage with your clients? If, for example, you are a personal injury lawyer, then Facebook may be a good marketing platform. If you have an intellectual property boutique, maybe potential clients find you through Google and you want a chatbot hosted on your website.

Choose a chatbot: There are several do-it-yourself platforms, such as Chatfuel, the relationship-based messenger marketing, Zendesk, the customer service and engagement platform, or Hubspot, for marketing, sales and service. When choosing a chatbot, consider the platform’s popularity (i.e., has it stood the test of time and will it continue to exist in the future?), the software’s ease of use, whether it is easy to integrate with your chosen messaging platform and, of course, the pricing.

Create a conversation map: You need to design a chatbot conversation, even if you work with a developer. It may be useful to do a legal process map and/or to create a decision tree that maps out the possible conversation paths. Note that though you may work with a developer, they will not be a substitute for the lawyer. Training a chatbot requires effort on the part of the lawyer just as much as the developer. Leaving the work solely in the hands of a developer or taking a throw-it-over-the-fence approach will not work. You are the expert and the person who will be able to formulate the right questions to best understand your client’s needs. If the questions are not skilful, several issues could occur: The chatbot conversation will feel unnatural to users, the client feedback may be poor because the app acts in a way to which clients may not be accustomed and the output may be factually unreliable because the chatbot failed to identify a legal issue.

Chatbots are just one example or instance of AI in the law. It does not include other examples in legal research such as ROSS or contract analysis such as Kira. The potential of chatbots is huge and AI generally, and its uses continue to grow. Hopefully, I have inspired you to experiment and create your own law firm chatbot.

Monica Goyal is a visiting professor at Osgoode Hall Law School. She is also a practising lawyer, an entrepreneur and tech innovator and founder of My Legal Briefcase.