AI agents are rapidly reshaping commerce, and organizations are making major investments to expand their capabilities and the scope of what they can buy, negotiate, approve, and execute. But hard questions begin where the transaction ends. What happens when an error occurs? What happens if harm is alleged? What happens when no one can clearly explain what an agent agreed to, on whose authority, under what terms, or under which jurisdiction?
Those questions are the focus of a new paper, co-authored by Bridget McCormack, president and CEO of the AAA-ICDR®, and David Hoffman, William A. Schnader Professor at Penn Carey Law, the University of Pennsylvania, titled “Agentic Commerce Needs Legal Infrastructure, and The Courts Are Coming.” The paper argues that the missing legal infrastructure around agentic commerce is not a problem likely to be solved through normal market evolution alone. Instead, it is a growing source of risk: when disputes arise, courts and regulators will address them using existing legal doctrines and tools that were not designed for autonomous AI agents, making outcomes likely to be costly, fragmented, and unpredictable. The paper also suggests that the better path is for the organizations building agentic commerce, together with the institutions supporting it, to help shape that infrastructure deliberately rather than leave it to be defined case by case through disputes and enforcement. The systems are already being deployed at greater scale and speed, while the legal infrastructure needed to support them remains underdeveloped.
Comparisons of agent-to-agent commerce to traditional e-commerce are imperfect. Traditional e-commerce became durable over time because it developed legal and operational structures that people could rely on, with human involvement central to consent, accountability, and enforcement. Agentic commerce, however, introduces a different environment — one that can operate at extraordinary speed, often with far less direct human involvement. That creates new challenges in determining consent, accountability, and what it takes to establish a trustworthy commercial record within agentic systems.
The paper explores why these issues need addressing now and what is at stake if the infrastructure for agentic commerce is shaped reactively through litigation and regulatory enforcement rather than deliberately by the organizations and institutions building this market.
Download the full paper to explore why legal infrastructure may be one of the most urgent and overlooked challenges in agentic commerce, and why the courts may play a defining role in what comes next.
This article is the second in a series on the missing legal infrastructure in agentic commerce. This first article, and its companion white paper on the Legal Context Protocol, can be found here.