What We're Hearing from Law Firm Leaders

At Uprise Partners, we've had the privilege of serving law firms for many years—from boutique practices to multi-office firms—and we've had a front-row seat to how technology is reshaping the legal profession. As AI adoption accelerates, we're hearing many of the same questions from managing partners, firm administrators, and IT leaders:

• How do we introduce AI responsibly?
• How do we protect client confidentiality?
• How do we measure whether AI is actually delivering value?
• How do we ensure attorneys are using approved tools rather than creating unnecessary risk?

AI Is Reshaping the Legal Industry

Every managing partner has heard the promise: AI will save associates hours on research, draft documents in minutes, and give the firm a competitive edge. In fact, legal professionals using generative AI roughly doubled in one year, jumping from about 31% to 69% according to the 2026 Legal Industry Report from 8am, that surveyed more than 1,300 legal professionals. Bloomberg Law's June 2026 State of Practice survey put individual AI use even higher, at 83%. And Thomson Reuters and Georgetown Law's 2026 State of the Legal Market report found legal tech spending grew 9.7% in 2025, the fastest pace the industry has recorded, with knowledge management tools growing even faster at 10.5%.

The Challenge Isn't Adoption—It's Governance

Despite this rapid growth, what's less clear is how legal organizations should govern AI adoption in practice, specifically how to know if AI is working, how to keep client data safe, and how to make sure the tools being used are the ones leadership has approved. The same 8am survey found 43% of firms don't have a formal AI policy, and only 9% have a written policy that's actively enforced. Fifty-four percent haven't yet rolled out formal training on responsible AI use.

Rather than approaching AI as another technology rollout, law firms should treat it as a business initiative that requires clear governance, measurable outcomes, and ongoing oversight. The following framework highlights the key questions every firm should answer to adopt AI securely, demonstrate its value, and ensure it supports both attorneys and clients.

Start With Defining Success

Before rolling out any tool, it helps to define what you're measuring. Useful metrics tend to fall into a few buckets:

• Time and throughput. Track the amount of hours saved on important tasks such as first-draft contract reviews, deposition summaries, or legal research memos. While time saved is a straightforward metric, it can be misleading if the AI-generated work requires significant review or correction.
• Quality and error rate. Track how often AI-assisted work required meaningful revision compared to work done through traditional workflows. This can be more telling than speed alone; the goal is accurate work delivered efficiently, not just faster drafts.
• Realization and billing impact. If AI reduces first-draft preparation from three hours to ninety minutes, the real question isn’t just how much time was saved, it’s whether that time is being reinvested into higher-value client work, faster turnaround, or increased capacity for billable hours.

These conversations are worth having early. Clio's2026 Legal Trends Report for Solo and Small Law Firms found that even though 75% of small firms report high AI adoption, only about 31–32% report an associated revenue increase with most saying it's too early to tell, or that their pricing models haven't yet caught up to the efficiency gains.

The upside for firms that get ahead of this is meaningful: the Thomson Reuters/Georgetown report found that firms with a documented AI strategy were 3.9 times more likely to report significant business benefits than those without one.

Where AI Creates the Most Value for Attorneys

The productivity case for AI in legal work tends to be strongest in a few specific areas, and focusing rollout energy there rather than spreading across every possible use case at once tends to produce the clearest early wins.

The aggregate numbers are encouraging: Wolters Kluwer's 2026 Future Ready Lawyer survey found 62% of legal professionals report saving between 6% and 20% of their work week through AI. A Forrester Total Economic Impact study commissioned by Thomson Reuters found a composite legal team realized $626,000 in increased productivity over three years. Thomson Reuters' Future of Professionals research estimates the average lawyer can free up approximately 240 hours per year through AI adoption—a value of roughly $19,000 annually at average compensation rates, rising to as much as $53,000 per lawyer annually for firms at the leading edge of adoption.

• Document review and due diligence. Large-volume contract review, redlining, and due diligence in transactional work tend to show some of the clearest time savings, because the tasks are structured and the outputs are verifiable.
• First-draft generation. Routine agreements, correspondence, and discovery responses benefit from AI-generated first drafts that attorneys then refine—shifting time from blank-page drafting to editing and judgment, which is generally a more rewarding and strategic use of senior attorney time.
• Knowledge management and matter research. Tools that can search a firm's own past work product, precedent documents, and institutional knowledge tend to deliver strong value by reducing time spent recreating work that already exists somewhere in the firm.
• Client-facing efficiency. Faster turnaround and clearer status updates can become a genuine relationship differentiator, particularly with sophisticated clients who are under pressure to manage outside counsel spend carefully.

The common thread: AI tends to help most with the parts of legal work that are necessary but not where the firm's deepest expertise lives. Protecting partner time for judgment, strategy, and client relationships—while letting AI handle more of the structural, repetitive work—is the productivity story that benefits both partners and clients.

Accuracy in Legal Research Tools: Build Verification Into the Workflow

As many in the legal industry may know, AI-assisted legal research has a known limitation: it can produce confident, accurately-formatted answers that cite cases that don't exist or misstate holdings. The scale of this challenge is well documented. Stanford RegLab and the Stanford Institute for Human-Centered AI tested leading legal research AI tools and found they produced incorrect information between 17% and 34% of the time, even in tools marketed specifically as hallucination-resistant for legal use.

The good news is that purpose-built legal AI tools performed substantially better than general-purpose chatbots—and building a simple verification step into standard workflow closes most of the remaining gap.

Practical guardrails:

• Require that any AI-generated citation be independently verified against a primary source (Westlaw, Lexis, or the court's own database) before it appears in a filing or client-facing document. Some legal AI research tools now build this verification in by linking every citation directly to the source document—a feature worth prioritizing when you’re evaluating vendors.
• Treat AI research output as a strong starting point for an associate's own research rather than a final answer, particularly on new or high-stakes questions.
• Document a verification step in matter files for AI-assisted research, similar to how firms document other quality control steps, so there's a clear record if a citation is ever questioned.

Protecting Client Data in the Age of AI

Law firm leadership has some unique considerations that go beyond what a typical business faces when adopting AI— privilege, confidentiality obligations, and malpractice exposure all shape what “responsible adoption” looks like in a legal context.

Imagine an associate copies excerpts from a confidential merger agreement into a free, consumer AI chatbot to summarize key provisions before a client meeting. If the firm has not approved the tool or verified how it stores and uses submitted data, that confidential information could be retained by the provider or handled in ways that conflict with the firm's confidentiality obligations.

Concerns like these help explain why the 8am survey found data security (46%), ethical issues (42%), and privilege concerns (39%) are the top reasons firms are proceeding thoughtfully on institutional AI adoption.

Know where the data goes. Before approving any tool, it's worth getting a clear answer to a few key questions:

• Is client information used to train the vendor's models?
• Is data retained after the session ends?
• Where is it hosted, and who at the vendor can access it?

The good news is that many enterprise AI agreements now include contractual commitments that client data won't be used for model training and will be deleted after a defined period. Getting those commitments in writing—not just referenced in a sales conversation—is a straightforward protection.

Turning Shadow IT Into an Opportunity

One of the most valuable things that firm leaders can do is create a clear, accessible path for attorneys and staff to use AI tools they find genuinely useful—because if that path doesn't exist, people tend to find their own. Kyle Peterson, Uprise Partner’s Sr. Director, vCIO recently hosted a webinar, Out of Sight, Out of Control: A Business Leader's Guide to Shadow IT, exploring exactly this dynamic.

Free consumer chatbots, browser extensions, and personal accounts on legal AI platforms have quietly become one of the more common sources of unmanaged data risk in firms of every size—not because people have bad intentions, but because they found something useful and had no clear way to get it sanctioned.

For example, an associate under a filing deadline may paste contract language into a free AI chatbot simply because it’s the fastest option available. Without a clear policy and approved alternative, well-intentioned productivity can quickly become an unintended confidentiality risk.

A few approaches that work well:

• Publish a clear, readable policy on what's approved, what's not recommended (such as pasting client documents into free consumer AI tools), and where to direct questions. Shorter, conversational policies tend to get read and followed; long, legalistic ones often don't.
• Make the approved path genuinely easy to use. If the sanctioned tool requires three approvals or is harder to access than a consumer alternative, some people will naturally route around it. Reducing friction for the recommended option is a meaningful security improvement.
• Use visibility tools rather than relying solely on self-reporting. Network-level visibility and vendor admin dashboards can give leadership a clearer picture of which tools are being accessed—not as a surveillance measure, but as a way to identify and address data risk before it becomes a problem.
• Create a fast, simple way for people to request evaluation of new tools. A lot of informal AI use starts because someone found something useful and had no easy way to get it properly reviewed. A standing intake process—even a simple form reviewed monthly—channels that energy productively and often surfaces genuinely valuable tools the firm can then adopt properly.

An AI Governance Checklist for Law Firms

A practical AI governance structure for a firm typically includes:

1. A short, clear, written AI use policy distributed firm-wide, with concrete examples of what's encouraged and what to avoid
2. A designated owner or committee responsible for vendor vetting, usage monitoring, and keeping the policy current
3. Defined success metrics tracked quarterly, starting with specific pilot practice areas before expanding firm-wide
4. Contractual data protection commitments from every AI vendor with access to client information
5. A standard verification step for AI-assisted legal research before it reaches a filing or client deliverable
6. A simple, fast intake process for evaluating new tools, so people always have a clear path forward
7. Periodic reporting to firm leadership on adoption, any risk findings, and productivity impact

None of this requires the firm to become a technology company. It's about applying the same disciplined, ownership-based approach that firms already bring to conflicts checks, ethical walls, and client data security—treating AI governance as a standing discipline rather than a one-time rollout decision.

For a related perspective on AI adoption across SMB organizations, I recently hosted this webinar: The 7 Biggest AI Mistakes SMB Leaders Make—and How to Fix Them. Many of the takeaways I discussed translate directly to law firm leadership.

If you'd like help thinking through AI governance for your firm or want to make sure your security infrastructure is built to support responsible adoption, we'd love to be a resource. Feel free to reach out to our team any time.