��������

Every large law firm now has an AI rollout plan; yet, fewer have a plan for what happens to professional judgment once the drafting is done in seconds instead of hours.

This gap sits at the center of conversations that , an executive coach at Apex Coaching, and , who leads learning and development at Reed Smith, have been having with their peers. Both collaborated recently at the legal industry’s conference to crowdsource ideas from talent management professionals around their observation that legal employers continue to treat AI itself as a fix while still struggling with the harder work of teaching people what AI-enablement means for them as professionals.

Why AI adoption is not solving the legal judgment problem?

Kelsch has spent months talking with chief talent officers and coaching clients about what firms are doing with AI, and what she has seen fits a pattern: Law firms have invested heavily in AI tool training, including teaching their associates which platform to use and how to prompt it, but firms still are struggling to figure out how to train lawyers in the judgment that will be critical as AI becomes more integrated into legal practice. “The law firms are not yet teaching the legal judgment piece,” Kelsch says. “AI output can't be the end of the process. There has to be that supervision and scrutiny of the AI output... We don't see many firms addressing that deeper layer yet.”

Part of the reason for this is that the problem is hard to structure. Kelsch argues that teaching someone to recognize when AI’s answer is wrong, incomplete, or subtly misaligned with a client’s goals requires a kind of judgment that traditionally developed slowly through years of researching, drafting, and handling matters themselves. Now, firms have to figure out how to teach that judgment when AI is changing the very process through which lawyers traditionally developed it.

What does good judgment look like when AI drafts in seconds?

To define what good legal judgment looks like, Reed Smith's Hakala says she takes guidance from the ideas of  two law professors — cognitive friction, from at New York Law School; and incidental learning from at Vanderbilt Law School:

• Cognitive friction describes the slow, frustrating process of creating a document by hand, making decisions line by line, getting stuck, consulting with colleagues, and eventually developing a feel for how the document, brief, memo or other work product can be structured.
• Incidental learning describes what a junior lawyer absorbs by reading through cases, precedent agreements, or disclosure documents that turn out to be irrelevant to the matter at hand but can quietly teach fact patterns and market practice along the way.

AI tools compress both processes into a fraction of the time by producing the same or better work product without the struggle that used to build understanding.

Hakala sees the necessity to frame judgment in the age of AI as a form of empathy through an understanding of how senior lawyers or clients will use the work and then calibrating accordingly.

For junior associates, “judgment is about assessing how you can be most helpful in the situation and how you fit into the project," Hakala explains. "When it comes to AI, judgment is about choosing the right tool for the task and knowing how much to validate the result before it becomes your own work."

Why associates are facing an identity crisis

A surprising and underappreciated problem unfolding among many young lawyers is the growing crisis of professional identity. In recent coaching sessions, Kelsch says she frequently encounters confusion among junior lawyers about what makes them valuable in the age of AI. She sees this as a recurring pattern, and adds that associates are receiving mixed messages.

“Some are being told by firm leadership to use AI to gain efficiency, but then the partners and clients they work with daily may be saying, ‘No, don’t use it... . We don't trust AI yet." Attorneys caught in this gap start wondering whether using the tools counts as cheating or whether avoiding them will leave them falling behind.

Hakala also sees an ongoing structural challenges around the billable hour that will continue to exacerbate this identity crisis. “For decades, a high billable number was not just a business metric. It was how many lawyers measured their own worth,” she says. Indeed, the billable hour was simultaneously how firms charged clients and a stand-in for lawyers’ professional worth. A lawyer who billed 2,200 hours a year could point to this number as proof of dedication, excellent performance, and value to the firm. AI now threatens to strip away this proof point without offering a replacement, which leaves many lawyers questioning their value or at least a way to validate it.

Ultimately, Hakala and Kelsch see a decoupling of the billable hour from lawyer performance, along with an overhaul of performance evaluation and compensation systems, as critical long-term changes. While these difficult structural changes take shape, Hakala and Kelsch each point to concrete interventions that law firms can pursue in the meantime:

• ‍Redesign competency models around new skills — Hakala says leaders need "to bring people together to think about what success will look like for the AI-enabled lawyer, given this professional identity shift that is happening." These questions are a good starting point to help define success.‍
• Build structured, hands-on verification processes — Kelsch offers a specific mechanism for developing judgment in the age of AI. Each associate should complete an assignment fully on their own without the use of AI, and then use an AI to do the same assignment. From there, let the associate figure out how to critique the AI output based on their own experience and guidance from their supervising attorneys. This approach intentionally recreates the necessary cognitive friction to build judgment.
• Introduce scenario-based, simulation training — Giving associates a simulated, realistic scenario and a safe, supportive, and low-stress opportunity to reason through how they would act in it is a key component of future lawyer development, says Hakala. Using AI-based simulations as a learning modality allows junior lawyers to make mistakes safely and get guidance from partners. Hakala also suggests supplementing this with , built from a real mentor's writing and interviews, that junior lawyers can query to practice difficult conversations before facing them in the real world.
• Rethink leadership development — Finally, Hakala argues that this moment requires law firms to invest in the development of their leaders, since these leaders are being asked to define new versions of what success and value look like for the profession.

For every large law firm that has an AI rollout plan, Kelsch and Hakala suggest that an equal amount of planning go into how they can develop judgment among their junior lawyers in an AI-driven environment that, on the surface at least, seems to take away much of the opportunity to do that.

You can register for