AI and Banking as a Warning for the Legal Profession 

Jake Heyer*

Last year, I wrote about how major law firms were using artificial intelligence (AI) and how this adoption could implicate different parts of the legal profession.¹ At the time, AI was mainly discussed as a tool to enhance productivity in areas such as research and drafting.² In the last year, however, developments in the banking industry have shown that this framing might understate the magnitude of the AI shift.³ Specifically, banking provides a concrete example of AI adoption changing from a tool for productivity enhancement to potentially changing the structure of the workforce.⁴ 

Banking and Structural Workforce Change 

Towards the end of last year, OpenAI hired over one hundred former investment bankers from firms including JPMorgan, Goldman Sachs, Morgan Stanley, Evercore, and KKR.⁵ The project behind the hiring is called Mercury. The premise of the project is to teach AI to build the financial models that investment banking analysts currently create.⁶ The AI model will receive instruction from a senior banker, create a proper Excel model for a leveraged buyout or restructuring, make changes based on feedback, and give a final product to the senior banker that is ready for client presentation.⁷ 

What makes this so different from previous AI projects is the ultimate goal.⁸  Instead of creating a system that assists or enhances productivity, the goal of the product is replacement.⁹ With all-in compensation for first-year investment banking analysts sitting between $170,000 and $190,000, investing in an AI system that can work faster, harder, and without error could save these large firms millions of dollars annually.¹⁰ Moreover, AI productivity is exponentially greater than that of any individual human.¹¹ A recent article describes an experiment in which six AI models competed head-to-head to generate Strengths, Weaknesses, Opportunities, and Threats (SWOT) analyses, a form of analysis typically included in pitch books or valuation reports.¹² In that experiment, the most advanced models required only ten to fifteen minutes to produce comprehensive SWOT analyses, while less sophisticated models were able to generate similar outputs in under a minute.¹³ Given the months of intensive, investment-banking-specific training that have occurred since this comparison, there is little doubt that AI models already demonstrate productivity far exceeding that of a human analyst and will continue to become even more productive over time.¹⁴ 

Why the AI Shift in Banking Matters for Law 

The significance of the AI shift in banking lies not in the industry itself, but in what it reveals about how AI changes professional labor once it moves beyond experimentation. Banking demonstrates that when AI becomes sufficiently reliable, firms may begin to redesign work around the technology rather than using it as a supplement. This matters for the legal profession because many of the structural features that made banking vulnerable to AI-driven job displacement are also present in law.¹⁵ 

At the junior level, legal work relies heavily on structured and repeatable tasks such as legal research, document review, contract analysis, and first draft writing.¹⁶ These tasks closely resemble the analytical workflows that AI systems are already being trained to perform effectively in banking. As discussed in my previous blog, law firms have already begun integrating AI tools into these areas, including platforms that assist with legal research, summarize large volumes of documents, and generate draft agreements.¹⁷ That same discussion anticipated broader adoption trends, and surveys now indicate that a majority of lawyers expect generative AI to significantly impact legal practice in the near future.¹⁸ 

The banking experience shows that efficiency gains do not remain confined to time savings.¹⁹ Once firms become confident that AI outputs are accurate and consistent, staffing decisions begin to change.²⁰ If AI systems can reliably handle early-stage analytical work, the economic rationale for maintaining large classes of junior professionals becomes less obvious.

Implications for Training and Firm Leadership

The most significant risk to the legal profession is not immediate job loss, but the long-term consequences of shrinking entry-level hiring. Junior associates historically develop professional judgment by performing repetitive analytical work under close supervision.²¹ If AI performs much of this work, firms may find it increasingly difficult to provide meaningful training opportunities while still meeting profitability expectations.²²

This dynamic presents a serious issue for partners and firm leadership.²³ Law firms rely on a steady pipeline of junior lawyers to develop future senior associates and partners.²⁴ If fewer juniors are hired today, firms may face leadership gaps years down the line, with fewer attorneys who have accumulated the institutional knowledge, judgment, and client experience necessary to assume senior roles. Unlike short-term efficiency gains, this problem compounds over time and is difficult to correct once it becomes visible. 

Banking illustrates how quickly this pipeline issue can emerge. Firms may benefit financially in the short run by reducing junior headcount, but the long-term effects on leadership development remain uncertain. For law firms, the lesson is not inevitability but foresight. Understanding how AI reshaped banking may help legal institutions anticipate and address similar challenges before they are forced to react. 

^*Jake Heyer, J.D. Candidate, University of St. Thomas School of Law Class of 2026 (Submissions Editor). 

1. Jake Heyer, Navigating the Future of Law: How AI is Reshaping Legal Practice, U. St. Thomas L. J. Blog (Jan. 29, 2025), https://ustlawjournal.blog/2025/01/29/navigating-the-future-of-law-how-ai-is-reshaping-legal-practice/#760260f6-5188-4af7-b897-f63bb54736fd-link [https://perma.cc/A2V9-55MP].  ↩︎
2. Id. ↩︎
3. Meridith Dennes, OpenAI Just Hired 100 Ex-Bankers to Replace You: Here’s the Timeline, Prospect Rock Partners (Oct. 21, 2025), https://prospectrockpartners.com/openai-just-hired-100-ex-bankers-to-replace-you-heres-the-timeline/ [https://perma.cc/CY3A-NE56]. ↩︎
4. Id. ↩︎
5. Id. ↩︎
6. Id. ↩︎
7. Id. ↩︎
8. Id. ↩︎
9. Id. ↩︎
10. See Investment Banking Analyst Salary Guide, Wall St. Prep (Nov. 19, 2024), https://www.wallstreetprep.com/knowledge/investment-banking-analyst-salary-guide/ [https://perma.cc/N9MX-W4S5].  ↩︎
11. See Michael Schopf, Outperformed by AI: Time to Replace Your Analyst?, CFA Inst. (June 23, 2025), https://blogs.cfainstitute.org/investor/2025/06/23/outperformed-by-ai-time-to-replace-your-analyst/ [https://perma.cc/JBP6-BKKW]. ↩︎
12. Id. ↩︎
13. Id. ↩︎
14. See id. ↩︎
15. See What Junior Associates Do, Simpson Thacher, https://www.stblaw.com/your-career/associate-life/what-junior-associates-do [https://perma.cc/94DA-P43K] (last visited Feb. 1, 2026). ↩︎
16. See id. ↩︎
17. Heyer, supra note 1. ↩︎
18. Heyer, supra note 1. ↩︎
19. See Dennes, supra note 3. ↩︎
20. See Dennes, supra note 3.  ↩︎
21. See How to Move From Associate to Partner: A Comprehensive Guide, Viele Consulting Grp. (Aug. 8, 2024), https://www.vieleconsulting.com/how-to-move-from-associate-to-partner-a-comprehensive-guide/ [https://perma.cc/WHA6-G8UG]. ↩︎
22. See id. ↩︎
23. See id. ↩︎
24. See id. ↩︎