Alexi has received independent validation for its legal intelligence platform through the results of the 2025 Vals Legal AI Report (VLAIR), confirming its performance in legal research accuracy, citation quality, speed, and reliability.
The VLAIR assessment evaluated several leading legal AI tools using a comprehensive benchmark designed to provide transparent and industry-wide comparisons. According to the report, Alexi scored 80% accuracy overall, surpassing the baseline lawyer accuracy of 71%. This represents a 9-point or 13% relative improvement compared to the manual efforts of legal professionals.
On questions involving single-jurisdiction legal issues, Alexi achieved a 79% accuracy rate, tying for the highest score among all platforms tested. The findings highlighted that Alexi demonstrated what the report describes as best-in-class reasoning for core legal topics assessed as part of the evaluation.
Increased scrutiny over AI in law
The report arrives amid increased attention to "AI hallucinations"-incorrect or fabricated outputs generated by artificial intelligence-in legal proceedings. The VLAIR team cited a 250% rise in such incidents in legal decisions since May. As a result, there is heightened scrutiny around the use of generative AI in legal research, making independent and data-driven benchmarking critical for industry stakeholders.
The Vals Legal AI Report was created to provide the industry's first independent, data-driven benchmark for evaluating legal AI tools, aiming to measure real-world performance, reliability, and compliance rather than relying on marketing claims or unverified assertions.
Citation quality and response metrics
The report highlighted Alexi's tendency to cite valid primary law 8% more often than ChatGPT, reducing compliance and malpractice risk that can arise from referencing secondary or invalid sources. Additionally, Alexi was ranked among the top three contributors for response speed, while maintaining a 99% response rate, contrasted with a four times higher non-response rate for ChatGPT.
"These results confirm what our customers experience every day: Alexi delivers the superior legal accuracy and compliance leading law firms need from AI, without compromising on speed or reliability," said Mark Doble, CEO of Alexi. "As we look ahead, this validation cements our role in defining how the next generation of lawyers will work with precision, speed, and trust at scale."
Impact for legal teams
Alexi's improved performance offers several operational advantages for law firms and legal departments. The company claims that the platform's higher accuracy reduces the need for lawyers to revise AI-generated outputs, thus cutting costs and freeing up time for other work. The increased likelihood of citing authoritative sources also aims to safeguard firms from risk, enhancing compliance and protecting reputations.
Integration with legal workflows is another area highlighted by VLAIR, with Alexi's high response speed and reliability supporting team-wide adoption. This, according to the evaluation, removes friction in daily legal research tasks and helps automate multi-step legal processes while retaining accuracy and valid citations across workflow stages.
Study methodology
The methodology behind VLAIR's evaluation involved testing participants, including Alexi and other AI platforms, on their ability to answer questions requiring reference to US federal and state laws, regulations, and judgments. The test set was created with input from Consortium Firms and legal experts to reflect authentic legal practice scenarios. The resulting 200 question-and-answer pairs covered diverse research tasks including confirming statutory definitions, interpreting factual patterns, identifying court rules, summarising case law, and multi-jurisdictional legal analysis. The scope of the dataset is described by VLAIR as one of the most comprehensive assessments of generative AI's capabilities for legal research.
With the latest VLAIR assessment, Alexi is recognised for offering accuracy and reliability in automated legal research and workflow tools, with independent data supporting its claims regarding citation authority and operational speed.
