New Open-Source AI System with 13 Agents Launched to Automate M&A Due Diligence

A new open-source system named "Due Diligence Agents" was released this month, offering a suite of 13 specialized artificial intelligence agents designed to automate and deepen the analysis of contracts during mergers and acquisitions. The system, introduced by developer Zohar Babin, aims to tackle the time-consuming and risk-laden process of manual document review by deploying AI to analyze legal, financial, and operational data within a deal's virtual data room. The launch addresses a long-standing pain point in M&A. Traditionally, deal teams must manually sift through thousands of documents under intense time pressure, a process that is both expensive and prone to error. According to industry analyses, this pressure often forces teams to review only a sample of 40% to 60% of available documents, creating significant risk of missing critical liabilities. The cost of senior analysts and outside counsel, often billing between $300 and $800 per hour for this work, further strains deal budgets. The emergence of sophisticated, open-source tools like this marks a significant shift for the M&A landscape. While large corporations have been adopting proprietary AI for years, accessible technology now puts powerful analytical capabilities within reach of small and mid-sized dealmakers. However, the output of any automated system is only as good as the strategic questions it's asked to answer. Babin's system is built on a multi-layered architecture designed for comprehensive, cross-domain analysis. The first layer uses a deterministic rule engine to identify dependencies between different business areas. For example, if the finance agent flags a potential revenue recognition issue, the system automatically triggers the legal agent to re-examine the associated contracts for enforceability and clawback clauses. This initial triage is based on programmed rules rather than a more resource-intensive large language model. The core of the system is its second layer, which consists of 13 distinct AI agents built using Anthropic's `claude-agent-sdk`. Nine of these are specialists focused on specific domains: Legal, Finance, Commercial, Product & Technology, Cybersecurity, HR, Tax, Regulatory, and ESG. After these specialists complete their parallel analysis, four meta-agents—a Coordinator, Summarizer, Reporter, and Cross-Referencer—synthesize the individual findings into a cohesive report. A third layer performs what the developer describes as "neurosymbolic cross-domain analysis," which cross-references all findings to uncover complex risks that might be missed in siloed reviews. This integrated approach is designed to solve the common problem where a clause that appears benign in one context represents a material risk when combined with information from another department. This project's launch reflects a broader industry trend toward what experts call "agentic AI." Unlike simple document search tools, AI agents are autonomous systems that can plan a sequence of tasks, execute them, and adjust their approach based on new information. Commercial platforms are already seeing widespread adoption. Harvey.ai, which provides AI tools for legal and professional services, reports that over 25,000 custom agents are operating on its platform for live transactions. Similarly, professional services giant PwC now uses such tools across its Deals practice to review more source material and surface risks earlier in the M&A process. For small and mid-sized companies engaging in transactions, this technology can dramatically level the playing field by accelerating the initial data review. In our experience, the greatest challenge is not just finding the data, but interpreting its strategic impact. An AI can flag a change-of-control clause, but it takes human expertise to model its financial exposure and negotiate terms accordingly. This is precisely where our mergers and acquisitions advisory practice adds value, translating AI-driven insights into actionable deal strategy. We guide clients through this entire process; to learn more, contact C&S Finance Group LLC at csfinancegroup.com. In practice, these AI systems are tasked with concrete, high-value extraction work. They can identify and pull specific provisions such as change of control triggers, assignment restrictions, termination rights, and indemnification caps from hundreds of agreements consistently. They can also be directed to analyze specific regulatory filings, like the Form ADV for an investment advisor acquisition, to detect compliance gaps or unusual fee structures. The objective is to automate between 40% and 70% of repetitive review tasks, ensuring 100% of documents are analyzed with uniform criteria. We see this as a critical evolution, not a replacement for expert counsel. The efficiency gains are undeniable, allowing deal teams to focus on high-level strategy rather than manual document sifting. Businesses preparing for a sale or acquisition should begin incorporating these capabilities into their planning, as they are rapidly becoming the standard for thorough and defensible due diligence. As both open-source projects like Due Diligence Agents and their commercial counterparts become more integrated into M&A workflows, the baseline expectation for diligence is set to rise. The competitive advantage will likely shift from the ability to simply conduct a comprehensive review to the sophistication of the cross-domain analysis and the strategic interpretation of AI-generated findings.