Claude Cowork's /legal:triage-nda Slays My First NDA Review—Is This the End of Billable Hours?

The Debut: First Encounter with Claude Cowork's /legal:triage-nda

The legal landscape is perpetually seeking efficiencies, but nothing quite prepared the industry for the debut performance of specialized AI tools aimed squarely at core, time-consuming functions. Sharing her experience on Feb 11, 2026 · 2:05 PM UTC, attorney @alliekmiller detailed her first "rodeo" using the Claude Cowork Legal plugin, specifically the /legal:triage-nda feature. This wasn't a theoretical exercise; it was a genuine test driven by necessity, reviewing a real Non-Disclosure Agreement (NDA) sent in by an active client. The context was immediate, the stakes were real, and the results, as we shall see, were paradigm-shifting.

This initial foray demonstrated the plugin’s capability to ingest complex legal documentation and apply a standardized (or customized) set of scrutiny parameters—a task that historically consumes junior associates' billable hours for days, or at least several valuable hours for seasoned counsel. The sheer act of delegating this foundational review to an LLM presented a moment of significant professional reckoning.

Performance Under Scrutiny: The NDA Triage Results

The core function of the /legal:triage-nda tool is remarkably straightforward yet deeply disruptive: it systematically reviews NDAs against established market standards or a law firm's proprietary criteria. It acts as a rapid-fire first pass, identifying areas of potential exposure or deviation from the norm that a human lawyer would meticulously seek out.

@alliekmiller’s immediate reaction was palpable: she was genuinely "rocked" by the tool's efficacy. The implication here is not just speed—though speed is certainly a factor—but the accuracy and comprehensiveness of the initial assessment. To have a complex document summarized, flagged, and risk-profiled almost instantaneously fundamentally alters the workflow calculus.

Specific Criteria Analysis

What separates a decent tool from a transformative one is the depth of its analysis. The AI didn't just confirm receipt; it prioritized and flagged specific elements of the NDA structure. While the exact metrics varied, successful triage typically centers on critical clauses such as:

• Definition of Confidential Information: Was it overly broad or suspiciously narrow?
• Term Length & Survival: Did the restrictive periods align with industry norms for the subject matter?
• Governing Law and Jurisdiction: Were these locations favorable or inconvenient for the client?
• Remedies and Indemnification: Were the potential financial liabilities appropriately capped or mutually balanced?

When an AI can accurately score these points, it means the human lawyer begins their review not at page one, but at page one with an executive summary of risks already prepared.

The Industry Tremor: Implications for Legal Practice

The immediate, visceral reaction to tools like this is encapsulated perfectly in the acknowledgment: "I can understand why legal stocks were rocked." This isn't just about administrative efficiency; it’s about the perceived erosion of value attached to foundational legal labor.

Efficiency vs. Expertise

The central conflict ignited by this technology is the efficiency versus expertise trade-off. Can an AI replicate the nuance of human judgment required when an NDA involves highly specialized IP or sensitive regulatory environments? Perhaps not perfectly yet. However, for the 80% of standard commercial NDAs, the AI performs the routine screening flawlessly and instantly. This frees the human expert to focus their billable time where judgment truly matters: strategic negotiation, complex structuring, and anticipating secondary consequences.

The challenge to the status quo is clear. Traditional legal workflow models, heavily reliant on junior staff logging time performing rote document review, are now directly threatened. Why pay for four hours of associate time to review a standard agreement when the AI provides 90% of the necessary analysis in four minutes?

The Billable Hour Reckoning

The most tangible consequence of widespread adoption of tools like /legal:triage-nda lies in the direct impact on low-level, high-volume tasks. Initial screening, first-pass markups, and comparison against template libraries are precisely the tasks that constitute the bedrock of the traditional billable hour model for many firms.

This inevitable efficiency forces a critical transition question upon the industry: Are we ready to move decisively away from time-based billing towards value-based pricing? If the AI handles the 'time,' the law firm must now package and sell the 'value'—the strategic insight derived after the AI has done its preliminary work. Charging for the initial document review becomes increasingly indefensible when the client knows a machine could have done it faster and cheaper.

A Call to Counsel: Lawyers Weigh In

This moment demands a collective response from the profession. @alliekmiller issued a direct challenge to her peers: "Lawyers, do we like this? Are you happy tiny tasks like NDA reviews won't hit your desk as often?" The mixed response is likely already forming across the profession—some fearful, others excited by the prospect of elevated work.

The future mandate for legal professionals hinges on adaptation. If the AI handles the triage, lawyers must pivot sharply toward specialization in areas where creativity, complex human negotiation, and deep regulatory understanding remain paramount. The role shifts from being a gatekeeper of information access to being a curator of high-level strategy derived from that information. The conversation is no longer theoretical; the tool is in the field. The call to action is clear for those seeking to stay ahead: engage with the technology, understand its limits, and prepare for the recalibration of what constitutes valuable legal work.

Source:

• Original Post: https://x.com/alliekmiller/status/2021586268573024337