Business functions are increasingly using generative AI to improve the quality and speed of operations. Savvy companies are incorporating AI tools in enterprise functions like sales, marketing, and engineering. AI now appears in many software tools, and vendors extol the benefits of AI.
Legal departments are being encouraged or required to incorporate AI solutions as well. But while much has been written about AI’s data concerns, regulatory issues, and hypothetical ethical issues, there is very little discussion of the practical applications of AI in the legal department.
Technologists frequently rank legal as one of the top areas where AI can assist. And while most people agree that it the wildest business-side dreams of replacing the legal department with AI tools are unlikely to materialize, the reality is that AI tools can be used to enhance the legal department’s productivity and operations. But which tools, and how should they be used? Which legal professionals should run the tools or have access to them?
Most important, which legal functions can realistically and reliably be performed, or enhanced, through an AI tool? Should we entrust AI with, say, drafting license agreements, or assembling legal processes? Can AI efficiently manage employee documentation processes? Should we base case settlements on AI analyses and recommendations?
The first step for any legal department is to have a clear understanding of the department’s processes. What kinds of work or legal projects does the handle, coordinate, or collaborate on? What volume of each kind of work flows through the department? How scalable do the chosen tools need to be?
Then the department can look into AI tools that are broadly applicable to the department’s processes. We suggest looking at broad applicability rather than having a narrow focus initially in order to learn more expansively about the capabilities or AI tools and offerings. Not only will this help the department apply more general information about tools and capabilities to its known processes to select the right AI tools, but also this broader approach may suggest productive applications of AI tools in the legal department that were not obvious or well-understood at the outset.
Armed with these insights, the department can then evaluate offerings from internal capabilities and/or external vendors. Any build-or-buy consideration will include the available budget and operational needs.
As one example, numerous businesses, such private equity firms, investment banks, and law firms, are employing AI tools such as large language models (LLMs) to draft and manage high volumes of nondisclosure agreements (NDAs) that they handle as a result of their high-deal-flow business operations. These businesses are finding that AI tools can be employed to review high volumes of NDAs, spot differences and flag anomalies, and use these findings to manage related legal obligations. However, in addition to these fairly standard-sounding uses of LLMs, some businesses are also using AI tools combined with data analytics to improve operational efficiency. For example, if a business works with a specific counterparty often, a trained LLM can generate a first draft of an NDA for that counterparty using the language that the counterpart has already agreed to in a majority of its previous NDAs. As another example, an LLM can be used as a rapid-feedback training tool that enables an attorney to request multiple versions of a clause or provision, with different shadings — say, language that favors one party or the other, or more neutral language. The potential applications of AI tools are potentially limited only by the creativity of the department.
The legal department should also consider the potential for AI tool usage to decrease the legal department’s value. A recent Boston Consulting Group article explored creative versus business problem-solving using AI, providing intriguing insights for Chief Legal Officers implementing AI. The studies suggest that while AI can improve overall performance, it might negatively impact the complex task performance of high-talent individuals. In other words, while average performance may improve with AI, high-performance output might suffer.
Departments and decision makers (General Counsel, Chief Legal Officers, etc.) implementing AI tools should consider what type of legal department they are building: an average one or one striving to be above average. Alternatively, they should consider which legal tasks could benefit from AI and which are better suited for human skills.