skip to Main Content
Aritificial Intellegence and Law Research

Can artificial intelligence make legal research more cost-effective without compromising quality?

Any practising litigator will know that thorough legal research is a precondition to success in almost every case. Litigators need to be up-to-date on the law, whether it be derived from statute, regulation, or jurisprudence, in order to properly advise their clients and effectively advocate for their positions.

The way in which legal research is conducted has evolved significantly in the past few decades with the advent of vast online directories such as Westlaw, LexisNexis, and CanLII. However, the process remains time-consuming, costly, and demanding, given the increasing complexity of the law and volume of information that must be gathered and synthesized.

To address these challenges, new legal software products have been developed which apply artificial intelligence and machine learning to the process of legal research. As a case study, we tested one recently-launched platform, Alexsei, against traditional research methods, to explore whether such products can make the process more efficient without compromising quality.

What is Alexsei?

Alexsei is a piece of legal software which seeks to provide high-quality answers to complex legal questions in a ready-to-use memo format.  The resulting memo is intended to be complete and ready for use. This is a tall order. Law is nuanced and asking lawyers to rely on a piece of software to provide an answer which recognizes and provides for those nuances is a challenge.  A quick review of Alexsei’s co- founders shows that one is a lawyer/software developer, one an ex-aerospace engineer with machine learning software expertise and one is completing his doctorate in artificial intelligence. We were very interested to see whether this combined expertise could actually deliver what is promised. If so, Alexsei has the potential to make a real impact in the daily life of a litigator.

The Contest

In order to put Alexsei through its paces, we devised two legal questions based on a unique set of facts. One lawyer would use traditional research methods to produce a research memo. Another lawyer would use Alexsei. The first question was a straightforward legal question:

1. Does the provision of free accomodation by a sibling over a 20+ year period constitute “providing support” under section 57(1) of the Succession Law Reform Act, for the purpose of determining dependants’ support payable by one sibling to the other?

The second question called for a more nuanced analysis, as follows:

2. Assuming a claimant is found to be a “dependant” of her testatrix sister under the Succession Law Reform Act, what quantum of dependant’s support can she expect if she’s been living in the testatrix’s home, free of charge, for 20 years?

The Results

Alexsei is very straightforward to use. There is a two-step process to inputting information; first you insert the legal question, then you insert the facts. Other than inputting the legal question and the facts separately, you are not required to enter the information in any particular format. This is one of the more impressive features of Alexsei. You can use natural language in the same way that you might discuss the issue with a colleague, or as the Alexsei website states,“[S]imply type in your question using natural language in the same way you’d email an associate”.  Prior to submitting your question, the user is required to input the jurisdiction of choice as well as any internal identifiers for billing or filing purposes.  The turnaround time is listed on the Alexsei website as 36 hours, however both memos were received within 24 hours.

In terms of presentation, the memos themselves appeared much like any other legal research memo. There was a “conclusion” section, which offered a summary of the analysis, there was a “law” section containing a detailed treatment of the law as well as an “authorities” section setting out the cases relied upon.  The text (surprisingly) appeared as if it had been written by a natural person and did not contain any clunky, disjointed paragraphs. The structure of the memo made logical sense and each point was addressed in a separate paragraph.  The Alexsei representative confirmed that every memo is a team effort between both artificial intelligence and the human lawyers at Alexsei, with the majority of the heavy lifting being performed by the A.I. algorithms. While an Alexsei lawyer does review the memo prior to delivery, this is primarily to ensure that the memo does not contain any grammatical or formatting errors.

The real question is: did the Alexsei memo produce an accurate legal memo? Given that there were two questions, we will deal with each question separately.

Question 1: In short, yes. Alexsei was asked to decide whether the provision of free accomodation, in the context of a dependant’s support application by a sister of the deceased, qualifies as “providing support”. Alexei stated as follows

No decisions could be identified where a sibling was seeking dependant support under the SLRA. However, other decisions discussing the provision of housing in the context of a dependant support claim may be instructive.

Effectively the same conclusion was reached using traditional research methods.

The Alexsei memo produced and summarized the relevant portions of a number of cases frequently cited by practitioners when discussing whether the provision of shelter satisfies the “providing support” requirement of the SLRA. Each case on the Alexsei memo was hyperlinked to the CanLII citation.

Notably, the memo created using Alexsei cited all the same cases as the memo created using traditional research methods, with the exception of one relevant authority cited by the Alexsei memo and not by the traditional memo. Thus, it is fair to say that the Alexsei memo outperformed the traditional memo in this respect.

Question 2: Alexsei did not provide a quantum of support based upon the inputted facts. What it did was it set out the factors contained at section 62 of the SLRA and provided the various quanta of support awarded in each of the cases discussed in the previous memo. This would provide a useful reference point for the lawyer considering this issue.

The traditional memo went through the same exercise, albeit in a more concise fashion. No specific conclusion on quantum was made by either memo, likely because the courts’ awards of dependant’s support are highly fact-dependent and notoriously difficult to predict accurately. Additional facts would have been required to come to a more specific conclusion.

Time Spent

The time spent for legal research and drafting using traditional methods was approximately four hours. In contrast, the time spent to input the research questions into Alexsei and review the memos it produced was only thirty minutes.

The Alexsei memos were available 24 hours after inputting each respective question.

Cost of Alexsei

The average memo ranges from $95 – $200, depending on the plan and the number of memos/questions per month. Alexsei is happy to customize each plan to the customer’s individual research needs.


Based on the above exercise, we are of the view that platforms such as Alexsei, which apply artificial intelligence to legal research, can be a useful tool in a litigator’s arsenal, and certainly if a lawyer is looking to delegate such work or does not have the in-house capacity to do the work. The answers to the questions posed in the contest we devised were comparable using Alexsei versus traditional research methods. However, Alexsei completed the task in much less time. This could result in significant costs savings for the client. We therefore encourage our colleagues in the litigation bar to consider how they can make use of this sort of technological innovation in their own practices.



The authors of this blog are Peter Askew and James Dunphy.

Related Posts and Articles
This blog is not intended to serve as a comprehensive treatment of the topic. It is not meant to be legal advice. Every case turns on its specific facts and it would be a mistake for the reader of this blog to conclude how it might impact on the reader’s case. Nothing replaces retaining a qualified, competent lawyer, well versed in this niche area of practice and getting some good legal advice.
Back To Top