Why AI Might Actually Create More Work for Lawyers

Episode Summary

Executive Summary: The episode examines how AI is reshaping legal work at a large law firm, with chair Gary Wingens arguing that AI is already improving legal quality, lowering costs on routine tasks, and changing training, pricing, and client expectations. The conversation highlights a likely Jevons-paradox effect: cheaper legal work may spur more demand, more matters, and more in-house automation rather than simply reducing firm revenue.

Main Topics: AI as a “co-pilot” for lawyers (Priority: 5/5): Wingens argues AI is not just automating repetitive work but improving judgment by serving as a thought partner that helps lawyers draft, test ideas, and produce better work product from the start. Billable hours and pricing pressure (Priority: 5/5): The hosts and Wingens discuss why billable hours persist, how AI is compressing the cost of specific tasks, and why hourly rates may still rise even as unit work gets cheaper. Training junior lawyers in the AI era (Priority: 5/5): The discussion centers on how law firms will need to replace grunt-work apprenticeship with simulator-style training and AI-assisted learning, while still preserving discipline and judgment. Security, privilege, and regulated adoption (Priority: 4/5): Wingens emphasizes that enterprise legal tools like Harvey matter because they provide secure, firm-controlled environments and reduce privilege and confidentiality risks versus consumer AI tools. Jevons paradox and legal demand expansion (Priority: 4/5): The conversation suggests lower costs may expand the volume of legal work—more litigation, more patent filings, and more internal review—rather than simply shrink the industry. Firm culture, knowledge sharing, and superstar lawyers (Priority: 4/5): AI depends on sharing expertise across the firm, which can clash with the autonomy of elite partners. The hosts probe whether superstar hires and proprietary knowledge will become even more valuable. Build vs. buy in legal tech (Priority: 3/5): Wingens says Lowenstein Sandler will customize existing tools and proprietary playbooks rather than train its own foundation model, citing cost, scale, and practicality.

Key Arguments: AI is materially changing legal practice, not merely speeding up old workflows; it can improve drafting and analysis by acting as a co-pilot. Routine, tedious work is the first to disappear, which should let junior lawyers move to more interesting work sooner, but raises training concerns. Alternative fee arrangements exist, but billable hours remain dominant because clients still want outside validation and law firms still monetize expertise that clients cannot replicate internally. Clients often bring AI-generated drafts to firms, but those drafts usually lack market nuance and transactional context. The legal market may experience a Jevons paradox: lower costs make previously uneconomic matters worth pursuing, increasing total demand for legal services. AI will likely shift some routine work in-house at large clients, but outside firms retain value through market breadth, judgment, and independent advice. Security and privilege are major adoption drivers; enterprise tools are preferred because consumer AI can create confidentiality and malpractice risks. Large firms will not train foundation models from scratch; instead, they will build customized layers and playbooks on top of frontier models. Even as AI reduces labor on tasks, hourly rates at top firms can rise because the billable hour becomes more productive and more valuable.

Data Points: Firm size: About 400 lawyers - Gary Wingens describes Lowenstein Sandler as a large New York-area firm. Rate increase at large firms: 10.1% - Wingens cites American Lawyer data showing 2025 average hourly rates rose by this amount at the largest U.S. firms. Rate increase vs CPI: ~3% CPI - He contrasts firm rate growth with overall inflation, suggesting AI may be boosting value per billable hour. Cost reduction on a diligence project: 70% lower - Wingens says an AI-assisted due diligence project reviewing thousands of trust agreements fell from an uneconomic quote to one the client accepted. Patent application improvement timeframe: Over 6 months - He says the firm had used AI tools to help draft patent applications for a little over six months and saw quality improve. AI usage cost in July: $36 - Wingens notes his own Claude use had consumed $36 so far that month. Model training cost: About $1.5 billion - He estimates training a foundation model from scratch would cost roughly this much, making in-house model training unrealistic for his firm. Kirkland AI investment: $500 million over 5 years - Used as an example of a major law firm investing heavily in internal AI infrastructure. Alternative fee adoption window: 15 years - Wingens says the move toward project-based and value-based pricing has been discussed for at least this long but has not broadly displaced hourly billing.

Pivotal Quotes: "I think it is a pretty huge change." — Gary Wingens: On whether AI is a major step change in how corporate lawyers work. "We are now at the point where their AI agent is talking to our AI agent." — Gary Wingens: Describing AI-to-AI interaction between clients and law firms in patent work and document review. "The future is a Jevons paradox for like administrative overhead." — Tracy Alloway: Closing reflection on whether AI will reduce bureaucracy or simply expand it.

Implications: AI is likely to reshape legal work by reducing routine labor, increasing demand for more legal services, and pushing firms toward secure, AI-native workflows. But it may also intensify pressure on training, firm culture, and pricing models rather than eliminating the billable hour outright.