The Context Crisis and the Myth of ‘Plug-and-Play’ Legal AI

By Nick Fleisher, CEO and Co-founder of Sandstone.

The modern General Counsel or Legal Ops Leader is caught between a rock and a hard place. On one hand, the mandate has never been clearer: move beyond being a cost center and become a strategic business partner. Boards want legal departments that can navigate geopolitical shifts, ESG mandates, and rapid-fire commercial expansion. On the other hand, the average in-house team is drowning in a rising tide of low-quality requests.

We were told that AI chat bots were the life raft. The prevailing narrative suggests that if you simply plug in a Large Language Model, your legal throughput will skyrocket. But for most departments, the reality is a frustrating stall. The barrier to legal effectiveness isn’t a lack of sophisticated algorithms; it is a fundamental business context crisis.

Unless AI tools are embedded directly into business systems and serve as workflow infrastructure, AI will remain a shiny tool that sits on top of a fragile foundation. True productivity gains come from end-to-end workflow intelligence.

The Anatomy of the Context Crisis

In most organizations, legal requests arrive as fragmented requests. A salesperson pings an attorney on Slack about a “standard” contract change; a marketing lead sends an email about a creative review; a product manager catches a lawyer in a hallway to discuss a feature launch.

In every one of these instances, the request is stripped of its commercial “reward.” The lawyer receives the legal question but is denied the business context. What is the deal value? Is this a strategic account we’ve been courting for years? What did we concede in the last three renewals?

This is the Context Crisis: the disconnect between the systems where the business lives (CRM, ERP, HRIS) and the legal work. When legal work is siloed in inboxes and document folders, the legal team is forced to operate in a vacuum.

The Reconstruction Tax: A Cultural Drain

When context is missing, attorneys pay what we call the Reconstruction Tax. This is the low-value time spent acting as a corporate detective before any actual legal analysis can begin. It involves chasing down the “why,” the “who,” and the “what happened before.”

The cost of this tax is not just measured in lost hours; the true damage is cultural and strategic:

• Default Conservatism: Human beings, especially attorneys, are naturally risk-averse when they are uncertain. When a lawyer lacks context, they cannot accurately weigh risk against reward. Consequently, they default to the safest “No” rather than finding the strategic “How.”
• Inconsistent Risk Tolerance: Without a unified view of the business, legal advice becomes idiosyncratic. One attorney might be aggressive on a deal while another is cautious, simply because they are working from different sets of informal context. This leads to a perception of Legal as an unpredictable bottleneck.
• Velocity Friction: In a world where speed is a competitive advantage, the Reconstruction Tax acts as a massive drag on company velocity. Business units, sensing the friction, may begin to make decisions without Legal’s input.

The value of a lawyer’s expertise is ultimately constrained by the quality of the information they receive. Even the most brilliant legal mind, when fed incomplete or distorted context, will respond slowly, hedge excessively, and optimize for safety over usefulness. Large language models behave the same way. When human input is vague, fragmented, or misaligned with the real problem, the model’s output reflects those constraints, producing cautious, generic, or inefficient responses. In both cases, the bottleneck is not intelligence, but the quality of the upstream input.

Beyond the Inbox: Building Context-Rich Operations

To move from a reactive “inbox model” to an AI-native department, the first step isn’t buying a point AI solution; it’s building a structured system of work. We must move institutional knowledge out of individual brains and fragmented threads and into an integrated legal intelligence that flows through every corner of the business.

A context-rich operation requires three things (in order of difficulty):

1. Connected Data: Data needs to be modeled across systems (e.g., making CRM data useful for sales contract reviews), mapped across relationships and time, and cleaned to get rid of noise vs. final versions. Legal departments need a bank of extracted terms and concepts that matter to them over time.
2. Codified Playbooks: Institutional knowledge must be digitized. Risk frameworks shouldn’t live in a PDF on a shared drive; they should be the logic that routes and informs every request.
3. Zero-Friction Intake: You cannot ask the business to change how they work. You must capture context from where they already live—Slack, Teams, or the CRM—and funnel it into a structured environment.

Legal tech should help to do this work. Legal workflows should be self-driving. Once the paths have been predefined and the context is accessible, AI’s full potential can be unlocked to drive productivity alongside in-house lawyers. This shift from system of record to system of work is what amplifies the capacity of a legal department.

Enter the AI-native legal department: powered by Unified Context

This is where the transition from “Traditional Legal” to “AI-Native Legal” becomes tangible. Solving the Context Crisis requires more than just a document repository; it requires a Unified Legal Context system.

2026 will be the year of AI-native legal departments. They will operate on the philosophy that AI is only useful if it is context-aware. These modern legal teams will adopt the next generation of legal tech which unified fragmented data and serves as the centralized home for the legal department.

Rather than expecting an attorney to jump between a CRM, a financial tool, and an old email thread, AI solutions should link these disparate data points directly to the legal workflow. When a request hits the legal team’s virtual desk, it is already “context-rich.”

By unifying data from HRIS, financials, and customer records, lawyers can ensure that every interaction is operating on a “perfect view” of the facts. This eliminates the Reconstruction Tax. Instead of hunting for information, the lawyer is presented with a synthesized brief that includes the legal query, the commercial stakes, and the historical precedent.

Conclusion: Context as the Competitive Differentiator

The path to becoming a strategic General Counsel is paved with information architecture. The “plug-and-play” myth of AI suggests that technology can replace the need for organized data. In reality, AI only amplifies the quality of the system it sits upon.

AI will not replace the lawyer. However, an AI-native system that provides a strong view of the business will amplify the lawyer’s judgment, allowing them to stop being detectives and start being the architects of the company’s success.

–

If you would like to learn more about Sandstone, please see here.

–

Author Bio:

Nick Fleisher is co-founder and CEO at Sandstone. An engineer by training, he spent the last several years leading the legal tech service line at McKinsey & Company in New York. At McKinsey, Nick’s focus was on AI & automation for law firms, corporate legal teams, and legal tech companies. Prior to McKinsey, Nick built over 10 mobile apps across AI and consumer. In his free time, he enjoys building niche AI agents and hacking things.

—

[ This is a sponsored thought leadership article by Sandstone for Artificial Lawyer. ]