Published: New England Law Update
Artificial intelligence is now a routine subject of discussion at law firms and throughout the legal profession. Clients are asking about it, firms are investing in it, and vendors are promising dramatic gains in efficiency. The debate is no longer centered on whether AI has a place in legal practice, but on how it can be leveraged most effectively. For transactional attorneys, the challenge is identifying the tasks AI can perform reliably and the areas where it remains an imperfect tool.
The reality is that AI is exceptionally good at certain tasks and considerably less effective at others. Understanding that distinction is becoming an increasingly important skill for transactional attorneys.
One of AI’s strongest transactional applications is document review during due diligence. Modern AI tools can quickly identify provisions relating to assignment, change of control, exclusivity, indemnification, and dozens of other issues across hundreds of agreements. Instead of manually searching through contracts one by one, attorneys can use AI to surface relevant provisions in multiple contracts all at once and in a fraction of the time. While results still require verification, AI can significantly reduce the amount of time spent locating information of value.
Similarly, AI is useful for locating information and concepts within a single document that may not be easily identified through traditional search methods. While a “Ctrl+F” search can locate a specific word or phrase, it generally cannot identify concepts or terms stated unconventionally. AI can identify concepts, provisions, and other relevant information based on the substance of the inquiry rather than a precise search term. This is especially valuable when relevant information may be buried in unexpected places.
AI has also proven particularly effective at drafting discrete contractual provisions. When provided with sufficient context, AI can generate initial drafts of routine provisions, suggest alternative formulations, and help attorneys tailor language to address a particular issue. This can be especially useful for provisions that follow established structures, such as confidentiality obligations, indemnification provisions, or other commonly negotiated terms. However, AI’s effectiveness in this context diminishes as the complexity or scope of the task increases.
Another area where AI performs particularly well is document comparison. Transactional lawyers are routinely required to compare a draft agreement against a previous version or another document entirely, like a term sheet or letter of intent. AI can identify, summarize, and categorize changes between documents, offering more insight and a more sophisticated analysis of revisions than a traditional redline, which merely displays textual differences.
While the benefits of AI in transactional practice are substantial, those benefits must be considered alongside its current limitations. Transactional practice often requires understanding why a provision exists, how it interacts with the rest of an agreement, what market expectations are for a particular deal, and whether a particular risk is worth accepting. These are questions that depend heavily on context, client objectives, negotiating leverage, business considerations, and experience. AI can provide useful observations, but it does not negotiate transactions, understand client relationships, or appreciate the practical considerations that drive legal strategy.
Similarly, AI remains unreliable when asked to make nuanced legal conclusions or strategic recommendations without careful attorney oversight. It may confidently overlook important exceptions, misunderstand how defined terms operate together, or fail to appreciate that a seemingly minor drafting change materially shifts risk between the parties. In many cases, AI’s analysis appears persuasive until a knowledgeable attorney reviews it closely.
Furthermore, AI generally performs best when given focused, discrete tasks. Asking it to “review this commercial lease and tell me if it’s market,” “draft a pro-buyer asset purchase agreement,” or “analyze this agreement from the seller’s perspective” is unlikely to generate useful work product.
Perhaps the most overlooked aspect of AI adoption is that “beneficial” does not always mean “efficient.” Many law firms understandably focus on whether AI reduces billable hours or increases attorney productivity. Yet using AI effectively often requires attorneys to formulate careful prompts, review every response critically, and ensure that the analysis aligns with a client’s specific transaction. The time spent reviewing and modifying AI-generated work can offset much of the initial time savings, particularly on sophisticated matters where precision is essential. But this certainly does not mean AI lacks value. Indeed, even where it does not materially reduce the total amount of attorney time, AI can improve the quality of that time.
As AI continues to evolve, so too will its role in transactional practice. The attorneys who benefit most are unlikely to be those who simply use AI the most, but those who use it with an understanding of its capabilities and limitations. Although AI will continue to improve, its most significant contribution today is not the automation of legal practice, it is allowing attorneys to spend less time locating, organizing, and processing information, and more time exercising the judgment, strategic thinking, and client counseling that remain at the heart of transactional practice.