This start-up has developed an algorithm capable of associating the requests of individuals with a legal situation, in order to recommend to them a list of relevant lawyers.
“A lot of people don’t know which lawyer to go to when they have a legal problem. Worse still, the complexity of the law and its jargon make it sometimes difficult to know which specialty to turn to. A public servant who has a problem at work with his boss may, for example, think that he should call in a labor lawyer. However, this concerns public law because he is a civil servant” explains Raphaël Jabol, co-founder of Avostart. This start-up solves the problem.
Created in 2016, the aspiring start-up developed an algorithm allowing anyone to find the most relevant lawyer to solve their problem. It claims 60,000 monthly users. Concretely, when a user asks a question, the algorithm translates his request expressed in natural language into a legal request. Thus allowing the user to get an initial answer.
The Avostart algorithm begins with a semantic analysis of a request expressed in natural language to extract a legal request. This is then crossed with an analysis of the sociological profile of the user, drawn up with different factors such as the city where he lives, the device on which he accesses the service, the time of access, or the age. These parameters can be used to evaluate the level of revenue for example with the type of smartphone (iPhone…) used to make the request.
It is thus possible to propose relevant lawyers according to the fees indicated by registering on Avostart.
Besides, the request is associated with other similar requests where the user has been satisfied with the connection with a particular lawyer: the lawyer in question will thus be recommended for a request of the same type. These steps make it possible to propose a list of relevant lawyers among the 55,000 listed in its directory.
AI at the service of lawyers
To translate a request expressed in everyday language into a legal request, Avostart used a machine learning model. This type of artificial intelligence research allows a machine to understand certain tasks by itself, training it to be able to do them on its own.
The start-up studied in depth the needs expressed by users and the relevant legal solutions that correspond before developing the app. About 30,000 requests enabled it to identify groups of legal needs. It is through this model that the algorithm understands how users express their demands and the recommended legal solutions.
Once the list of relevant lawyers is recommended, they are notified by e-mail.
If the user wants to go further, he can for 39 dollars get a 20-minute telephone consultation with the lawyer who provided him with a response. However, the lawyer in question must pay a subscription fee. This is the case with “several hundred” lawyers according to Avostart, which provides legal consultants with access to paid files. In other words, the people asking for consultation are clients who are going to pay the lawyer, which is not always the case with the free question-and-answer system. On the other hand, when the lawyer does not pay a subscription, the user is obliged to call another lawyer from the list of those recommended for consultation.
The relevance of the algorithm
Moreover, lawyers are not always the ones who respond to requests. When the algorithm analyzes issues requiring only legal information such as legal texts, and not legal advice, it then provides an automatic answer.
Finally, the user is sometimes contacted by Avostart after making a request, in order to improve the relevance of the recommendations of the algorithm. In other words, the user must clarify his request to an advisor of the start-up who is capable of understanding it since the algorithm can sometimes fail to identify the need and recommend a list of relevant lawyers.