#21 Can AI Replace a Junior M&A Associate?
This week, I discuss whether AI can act like a junior associate in a law firm and a feature of Microsoft’s new AI laptops, Recall.
AI will change the world, but how will it change M&A? I want to focus on AI’s impact on M&A in this newsletter. I am not an expert on either M&A or AI, but I want to learn about both topics and how they intersect. I thought there might be others in my situation (or people who are experts in one field or the other) who would find information on M&A and AI helpful in their careers, so I created this newsletter to track and share what I learn.
Can AI Work like an Associate in a Law Firm
Last week, a reader asked whether AI could be prompted to act like a junior M&A associate. This week, I want to dive deeper to try to answer that question.
The Task
First, it’s important to outline what the AI must do to match the capabilities of a junior associate in a law firm.
I am not an associate at a law firm (yet), but I think I have a basic understanding of what goes on.
For purposes of this post, let’s assume the junior attorney’s task is to take the first stab at drafting an agreement based on precedent from a different deal found in the firm’s document management system.
Expectation
Second, we need a standard for the AI to meet. I think the general expectation of a junior associate from more senior lawyers and partners is fairly low—a few substantive mistakes in the document are expected. The associate should get all of the little things correct, like formatting, substantive provisions on which the associate received specific instructions, and material that the associate has seen on other agreements. The AI tool or technique must meet this requirement to “act like an associate”
AI Tools or Techniques that Can Complete The Task and Meet the Expectations
Prompting:
One technique for achieving this goal is separating the agreement into various parts and prompting a chatbot, like ChatGPT, to change the specific provision to match the client’s needs. I think this technique works for simple provisions, like the choice of law provision where the language is fairly common throughout all agreements. Making a substantive change would be much more difficult, requiring either a very detailed prompt (in which case why not write out the provision yourself?) or a very close precedent provision that would give ChatGPT a good template for how to draft the current provision. Needless to say, when drafting requires substantive or unique changes, prompting is a less-than-ideal replacement for an associate.
Spellbook has a couple of key features that could help it act like a junior associate. First, it allows a user to mark up a document by prompting a GPT-4 connected model. Spellbook then will redline a document, leaving comments where it makes changes. This feature is the closest thing I have seen an AI tool get to making substantive changes to a document. Second, Spellbook can “repurpose a contract for a new deal.” The user can prompt the system to make changes to an entire document. In the video, the prompt asked the AI to change the parties and the choice of law provision. These are obviously easy changes to make. Something that I would be interested in seeing is whether this Spellbook could make more substantive changes to the document based on a client’s goals or strategy. I don’t have enough information to determine whether it could handle such a request, but I think it is safe to say it would be difficult.
AI Agents
Agentic technology does not currently exist in a form that could complete the task, but I think AI agents could eventually draft an agreement like a junior associate. For a basic understanding of AI agents, here is a previous post on the topic. An agentic AI model would send various AI models who are trained to be experts on a specific task to draft a document. In this case, the models would be trained to be experts on various provisions in an agreement. The advantage of an agentic model is that one prompt would send the agents to draft the agreement and complete the task that would currently take several prompts. The downside is that the prompt would have to be very specific and give extreme detail on any substantive changes that are needed to the document.
Conclusion
Out of the three possibilities that I highlighted above, I am not certain that any of them could act like an M&A associate at a law firm. I think all of the options are (or will be) adequate for completing simple tasks, but all fall short when asked to make more substantive changes. If anyone has any insight on other tools or techniques please let me know!
Microsoft’s “Recall”
Microsoft recently released a forthcoming feature of new Copilot-enhanced laptops called Recall. Recall takes screenshots of everything the user does and stores them locally on the computer. The user can then use AI to search for anything from photos viewed to emails read using natural language. For example, a user could prompt Copilot to “find the email I wrote to John about dinner tomorrow night” or “pull up the Wall Street Journal article about Bitcoin I read yesterday.”
How will this affect M&A? The most immediate impact I think will be enhanced search, like the example above. In the long term, I think this technology will help M&A lawyers analyze and make good decisions regarding dealmaking. Recall, and related technology, could act like a set of eyes over the shoulder of the M&A lawyer, double-checking the decision-making of the lawyer.
Here is how I think Recall combined with AI agents1 could have resolved the SaveMart debacle:2
Recall could track the relevant definitions used by the law firm;
Recall or an AI agent would notice that the SaveMart debt defintion differed from the defintion typically used. Alternatively, Recall and an AI agent could “see” the JV debt guaranteed by SaveMart and “see” that the debt definition included the JV debt; and
An AI agent could warn a lawyer of a potential mistake in the definitions.
Although it is a little creepy (an AI taking screenshots and seeing and reading everything you do on your computer), Recall and other new developments like AI agents could potentially lead to more efficient and accurate lawyering. I think this is just the beginning of AI “watching” what we do on our computers, and it is yet to be seen whether people are comfortable with that level of intrusion. If it catches on, I think this technology will be helpful, especially when paired with AI Agents.
About me
My name is Parker Lawter, and I am a law student pursuing a career as an M&A lawyer. I am in my last semester of law school, and with some extra time on my hands, I decided to create this newsletter. I hope it is informative and helpful to anyone who reads it! I am not an expert at either M&A or AI, but I am actively pursuing knowledge in both areas, and this newsletter is a part of that pursuit. I hope you’ll join me!
Follow me on LinkedIn: www.linkedin.com/in/parker-w-lawter-58a6a41b
All views expressed are my own!