Law
GenAI
December 6, 2025
The legal technology landscape is evolving rapidly, with innovations aimed at streamlining processes and enhancing efficiency within legal firms and corporate legal departments. A noteworthy advancement is the launch of MARC, a generative AI-powered document review system developed by Altorney. This system is designed to automate first-pass review decisions, significantly optimizing the e-discovery workflow. After a successful pilot phase involving corporate legal departments, MARC is now widely available for use by various legal entities, including litigation service providers and law firms, thereby targeting substantial cost savings in the e-discovery process.
MARC addresses a pervasive inefficiency within e-discovery workflows: the tendency of organizations to upload extensive document sets into costly review platforms, only to eliminate a vast majority as non-responsive. As articulated by Shimmy Messing, CEO and co-founder of Altorney, this method not only incurs unnecessary expenses but also poses significant security risks. By automating the initial culling and review decisions prior to documents reaching the review platform, MARC ensures that only relevant documents—those already tagged with initial assessments regarding privilege, confidentiality, and responsiveness—are uploaded, thereby mitigating potential risks and costs.
MARC functions as a text analytics tool positioned between data collection and the review platform. Notably, it is designed to be agnostic concerning the large language models (LLMs) it can utilize. Organizations have the flexibility to deploy MARC with Altorney’s proprietary Llama model installed locally or integrate it with their preferred models from providers like Azure or OpenAI. This configuration allows all data to remain within the organization’s firewall, thereby enhancing security and reducing costs associated with cloud-based AI services.
A distinctive feature of MARC is its avoidance of user-dependent prompt engineering. Instead, it employs a protocol analysis method, allowing users to upload background materials related to their cases. MARC generates a detailed protocol document that can be edited by attorneys to refine the relevance parameters for the analysis. This approach maintains familiarity for legal professionals, enabling them to edit documents in Microsoft Word without requiring expertise in prompt engineering.
Once the protocol is established, MARC can integrate data from various sources, including text files and databases. The system employs a sampling and validation workflow to verify its results, ensuring that only statistically valid sample sizes are analyzed and tagged as relevant or non-relevant. This iterative process continues until legal teams are satisfied with MARC’s performance, allowing for processing speeds exceeding one million documents every 24 hours.
Beyond simple relevance assessments, MARC is capable of conducting multiple analyses in a single pass, including privilege review, personally identifiable information (PII) detection, issue coding, confidentiality analysis, and foreign language processing. These capabilities enhance the system’s utility, providing legal teams with comprehensive insights into the documents being reviewed.
MARC not only delivers decisions but also explains its reasoning for each determination, which is vital for maintaining transparency and defensibility in legal contexts. This feature empowers legal teams to scrutinize and understand the AI’s decision-making processes, facilitating informed adjustments to protocols as necessary.
In pilot testing, MARC demonstrated significant efficiency improvements, with one Fortune 500 client’s review costs reduced by 62%, and hosting costs by 78%. Furthermore, MARC’s cost estimations proved remarkably accurate, showcasing its potential for predictable budgeting in AI-driven e-discovery processes. This predictability can enhance the budgeting process for legal departments, which often struggle with fluctuating costs associated with traditional review methods.
The introduction of AI-driven systems like MARC heralds a transformative shift in the legal industry. As artificial intelligence continues to evolve, the capabilities of tools like MARC are likely to expand, offering even more sophisticated analyses and efficiencies in document review. Furthermore, the integration of AI in legal workflows is expected to redefine the roles of legal professionals, focusing their expertise on high-value tasks while allowing AI to handle more routine processes. This shift could lead to increased productivity and improved outcomes for legal teams, as they leverage technology to enhance their practice.
Disclaimer
The content on this site is generated using AI technology that analyzes publicly available blog posts to extract and present key takeaways. We do not own, endorse, or claim intellectual property rights to the original blog content. Full credit is given to original authors and sources where applicable. Our summaries are intended solely for informational and educational purposes, offering AI-generated insights in a condensed format. They are not meant to substitute or replicate the full context of the original material. If you are a content owner and wish to request changes or removal, please contact us directly.
Source link :
