逾 7,000 名法律界人士談 AI 需求：更多互不連通的工具並非答案

By Brent Homan, vice president, UX design, LexisNexis Legal & Professional

It's late, and the motion is due in the morning. The research was done in one AI tool, and the first draft was written in another. Firm precedent is buried in a separate system, and redlines are in Word. For a lawyer preparing a motion, each switch between tools means re-explaining the matter, cleaning up output, and rechecking the authorities.

AI was supposed to free lawyers to focus on the work that matters most: sharpening strategy, advancing decision-making, honing client counsel, and applying their expertise. Instead, managing disconnected tools has become part of the job.

Lawyers describe consistent AI frustrations

In 2026, the world's largest AI companies launched dedicated legal offerings, while leading legal AI startups have raised more than $1 billion in funding. When LexisNexis surveyed legal professionals a year apart, the average number of AI tools used per person rose from 2.84 to 3.34 — from less than three tools in 2025 to more than three in 2026.¹

The question is no longer whether to use legal AI. It's how many tools to juggle and what happens in the gaps between them.

This year, LexisNexis conducted over 170 user studies and focus groups on AI, involving over 7,000 legal professionals from Am Law 10 firms, in-house legal departments, and solo practitioners.² Across these studies, lawyers described a frustrating pattern of moving between AI tools to finish a single matter, recreating context, reformatting generated drafts, and rechecking citations before relying on them.

Whatever tools are used to prepare a draft, the lawyer must assess the argument and the authorities supporting it. Using multiple tools also means repeating instructions, reformatting, and tracing claims back to sources.

From AI friction to connected, verifiable work

At the Customer Innovation Lab, where LexisNexis conducts in-person research and builds live with customers, we ask legal professionals what would make AI work for them. They don't want another tool. They want AI that keeps up with the matter, minimizes rework, and makes sources easy to verify. That means:

• Context that carries through. The matter's facts, documents, and authority remain linked throughout the work, from research to review. Returning to a draft or handing it to a colleague doesn't mean starting over.
• Drafts ready for review. Work arrives with its styles, numbering, footnotes, and tracked changes intact, so review focuses on the argument, not the formatting.
• Verification built into the work. Sources stay visible and linked, so a lawyer can check whether authority supports a claim. The judgment and responsibility stay with the lawyer.

For LexisNexis, this thinking shaped the latest expansion of Lexis+ with Protégé to bring legal authority, matter context, drafting, analysis, news, verification, and the organization's prior work into a single prompt box.

For the lawyer on that motion, a single request can draw on authoritative case law, docket records, analytics on the assigned judge, legal news, and the firm's prior documents, each linked to its source. Shepard's provides the lawyer with a built-in check on citation status and related quotations. The same principles apply when in-house counsel reviews an agreement or conducts diligence.

Considerations for leaders

For legal leaders, the studies suggest a new evaluation framework for legal AI focused on how the tool contributes to finished, review-ready work. When evaluating AI investments, ask:

• How much value gets lost in the handoffs? Trace a matter from the first question to the final draft, and count each time a lawyer has to re-explain, re-upload, or reformat.
• Can lawyers check whether the sources support the work? Legal AI should help lawyers assess a case's status and treatment at any stage.
• What does it know that a general-purpose model doesn't? Authoritative content, citator data, and the organization's knowledge separate trusted legal work from plausible-sounding text.
• Does the output arrive ready for review or ready for cleanup? Style and formatting rework is easy to miss when evaluating AI, but at scale, it drains productivity.

For the lawyer finishing that motion, the argument and its supporting authorities still need careful review. Legal leaders should consider how much coordination tools require before that review can begin.

Help build the future of connected legal AI: Join LexisNexis at the Customer Innovation Lab in New York.

This sponsored post was supplied by LexisNexis.

_{1. LexisNexis longitudinal AI-tools survey of Lexis+ with Protégé subscribers, n = 276, baseline August 2025, follow-up August 2026.}

_{2. LexisNexis Press Release, "As Lawyers Face AI Switching Overload, Lexis+ with Protégé Redefines Integrated, Verified Legal Work — in a Single Prompt Box" (October 27, 2026).}

• AI
• Studios Enterprise