Of all theconsiderations that go into building Alexsei, memo quality is of supremeimportance. In fact, this is the single metric on which most of our time,energy, and resources are focused. Without high-quality answers, the otherbenefits of using artificial intelligence—such as reduced cost and turnaroundtimes—are pointless.
While there are many factors that influence memo quality, there is one that is largely out of our control: how the legal question is phrased. Of course, a single legal question can be asked in many ways. And while Alexsei is designed to account for this, sometimes a question phrased one way results in a higher quality memo than a question phrased another way. As such, we decided to put together this short guide to asking good questions when using Alexsei.
There is only 1 rule that must be followed when a user asks Alexsei a question. If it isn’t followed, the question will be returned and marked as unanswerable. There are also 4 guidelines that should be followed to maximize memo quality, although compliance with these guidelines isn’t necessary.
As a quickreference, here is the rule and guidelines:
Rarely, buton occasion, a user may ask a question that is either a best practices questionor a question that is more of a business question than a legal question. Unfortunately,Alexsei does not answer these sorts of questions.
As a rule of thumb, if you believe that reference to source law is entirely sufficient to answer your question, you are probably right, and your question is unlikely to be returned.
When submitting a legal question to Alexsei, users are provided with a form field for the legal question and a form field for any background facts. Often, users include both facts and legal questions as a single entry in the question form field. In these cases, there is a risk that the answer will not be as helpful as it would have been had the user properly separated out the facts from the issue statement.
This istrue for two reasons.
First, if usersdon’t properly separate out facts from the legal question, we are often forcedto do it on the backend manually before running the question through our AI, andtherefore risk misinterpreting the precise legal question.
Second, the law is inherently objective—it is an abstraction that is broadly applicable to certain facts. Facts, on the other hand, are specific to a subject. Accordingly, facts and law consist of distinct domains, although related. As such, when an issue statement is provided distinct from facts, properly identifying the most applicable source law become significantly easier.
As a ruleof thumb, the level of generality or specificity of an Alexsei answer willcorrespond to the level of generality or specificity of the question.
When a question is intentionally general or intentionally specific, the best answer will likely be helpful. However, when the question is unintentionally general, even the best answer will likely fall short of what the lawyer was looking for. Formulating a specific legal question does require some general knowledge of the area of law, which, on occasion, requires some work up-front. One solution to this, however, is to start out by intentionally asking Alexsei a general question (Seen #3 below).
As previously noted, specific questions are more likely to result in a helpful memo than general questions. However, specific questions are not always possible for several reasons, one of which is a lack of understanding of the area of law. In these cases, intentionally asking a general question is entirely appropriate, and will provide a good base from which more specific questions can be formulated. In fact, we often see a series of questions from a user follow this pattern—progressing from the general to the specific—reflecting the natural development of the user’s understanding of the relevant law.
As alllegal professionals know, precision of language is critical. Not surprisingly,this still holds true when asking Alexsei a legal question. For example, if auser has a question about breach of contract, and uses the word “rescission” inthe question when they really meant “repudiation,” the answer returned is goingto be significantly different than what they were expecting (not surprisingly!).In some instances, being precise in yourquestions may require some up-front work. Most of the time, however, precisioncan be achieved simply by paying close attention to your question, and doublechecking the meaning of any obscure words or phrases where possible.