What We Learnt About AI Adoption on Real Legal Matters at NAMWOLF

October 30, 2026

As a sponsor of the 2026 NAMWOLF Annual Meeting, Cimplifi was proud to host and moderate a CLE session titled AI in Practice: What We Learnt on a Real Matter. The goal was intentionally practical. We weren’t there to debate whether AI will change legal work. We focused on what happens when lawyers and legal departments actually use AI on real matters, with real clients, real risk, and real accountability.

The panel brought together three perspectives that don’t often share the same stage: law firms actively deploying AI, firms still building their strategy, and corporate legal teams evaluating what they’ll approve and trust. Joining me were Ajay Mago of Maxson Mago & Macaulay LLP, Brian Harrell of Quintairos, Prieto, Wood & Boyer, and Eleni Katsoulis of AbbVie.

What made the conversation especially valuable was its honesty. The discussion wasn’t about replacing lawyers with AI. It was about identifying where AI creates value, where it introduces risk, and what legal professionals need to do to use it responsibly.

The Three Barriers Holding Most Legal Teams Back

One of the central themes of the session was that most legal teams don’t struggle with AI because they lack access to tools. They struggle because they encounter one of three barriers and stop there.

The panel framed legal AI use cases in three categories:

  • Everyday work such as drafting, summarising, and outlining.
  • Volume work such as document review and large-scale analysis.
  • Judgement work such as privilege review and case strategy.

Each category tends to stall for a different reason:

  • Everyday work is often blocked by concerns around client confidentiality.
  • Volume work is often blocked by economics.
  • Judgement work is often blocked by oversight and defensibility.

That framework resonated because it reflected what many attendees are experiencing today. Most organisations have experimented with AI. Fewer have moved beyond their first successful use case.

Confidentiality Is Still the First Gate

When legal teams begin exploring AI tools, the first challenge is rarely technical. It is understanding how client information is handled and whether it can be used safely.

The discussion highlighted a critical distinction between consumer-grade AI tools and enterprise-grade implementations. Two versions of what appears to be the same product may handle data very differently, particularly regarding retention, training, and access controls.

The panel also spent time discussing California SB 574, which becomes effective January 1, 2027. The law requires attorneys to understand how AI systems handle client information, take reasonable steps to verify AI-generated output, and personally verify citations used in court filings.

The larger takeaway was that AI governance is no longer simply a best practice. Increasingly, it is becoming part of an attorney’s professional responsibility.

AI Isn’t Replacing Review. It’s Changing It.

Document review generated one of the most practical discussions of the session.

For years, legal teams have relied on technologies such as deduplication, analytics, threading, and TAR to manage growing data volumes. AI-assisted review adds another layer, helping teams perform criteria-based coding, issue identification, and summarization at scale. But the foundation remains the same: human validation.

One point echoed throughout the panel: supervision does not transfer to software. Teams still need documented protocols, sampling strategies, quality control workflows, and defensibility measures. AI can accelerate review, but it does not eliminate the obligation to validate outcomes.

The panel also emphasized a lesson many organisations learn the hard way: not all data behaves the same. Non-English content, formula-heavy spreadsheets, and image-rich files may require different handling and validation approaches. Understanding your data before deploying AI is essential.

The Economics Question Is Real

One of the most candid parts of the conversation centered on economics.

For many law firms, the challenge isn’t whether AI works. It’s who benefits from the efficiency. On hourly matters, clients often capture the savings while firms absorb the cost of implementation, training, and process redesign. The incentives look very different under alternative fee arrangements, flat-fee engagements, or contingency matters.

The panelists discussed approaches that have helped organisations move forward, including per-matter pricing models, treating initial deployments as investments rather than immediate cost-saving exercises, and discussing AI use during budgeting conversations rather than after work has already been completed.

“I believe if you can deliver value, no matter what you do, then your customer is going to recognise that. The costs are going down, but the delivery of value is going up.” — Ajay Mago

Perhaps the most important observation was that AI adoption is not merely a technology decision. It is a business model decision. Firms must determine how they will capture value from greater efficiency while continuing to deliver exceptional client service.

High Value Means High Oversight

When the conversation shifted to privilege review, early case assessment, and case strategy, the topic became less about technology and more about trust.

The panel discussed how AI can help surface potentially privileged documents, create chronology summaries, identify key issues, and uncover patterns that attorneys might otherwise miss. Yet these are also among the highest-risk legal decisions an organisation makes.

The consensus was clear: effective oversight is not a feeling. It is a process.

“My number one favourite phrase in life is human in the loop. Validate, validate, validate. AI tools should be a thought partner. They don’t replace your judgement.” — Eleni Katsoulis

That process includes defining criteria before the tool runs, establishing sampling methodologies in advance, conducting secondary reviews, and documenting decisions contemporaneously. These controls help transform AI from an interesting experiment into a defensible legal workflow.

The discussion also reinforced a theme that surfaced repeatedly throughout the session: AI can enhance legal work, but it does not replace professional judgement.. Attorneys remain responsible for the outcomes.

What We Learnt from Real Matters

The most engaging portion of the session was when each panelist walked through actual use cases and lessons learnt.

Brian Harrell described how he is using AI in litigation:

“What I find the most value in, at least in my litigation practice, is I use it as a thought partner. You can literally cross-examine an expert virtually, and it will help you develop the outline you’re going to use to cross-examine the actual expert.” — Brian Harrell

The conversation focused on three questions: What did we do? What happened? What stopped us from going further?

Across those examples, several common themes emerged:

  • AI delivers the most immediate value when applied to repetitive, high-volume tasks.
  • Verification remains a critical part of every workflow.
  • Governance and training matter as much as the technology itself.
  • Internal champions help organisations move from experimentation to adoption.
  • Firms that establish guardrails early are better positioned to scale successful use cases.

Interestingly, some of the most valuable lessons came from discussing limitations rather than successes. Understanding what prevented broader adoption often proved just as instructive as hearing about successful implementations.

Where Legal Teams Should Start

As we closed the session, we focused on a practical question: What should legal teams do next?

The recommendation was straightforward: start with work that is high volume, low stakes, and easy to verify. Run the process both with and without AI. Document what works. Create written policies. Train users. Then expand from there.

We also encouraged attendees to ask every AI vender six critical questions:

  • Where does my data go?
  • Is it used for training?
  • How long is it retained?
  • Can I see the sources?
  • How is output validated?
  • What happens if I leave?

Final Thoughts

The biggest takeaway from moderating this discussion was that the conversation around legal AI has matured.

A year or two ago, most discussions focused on the technology itself. Today, legal teams are asking more practical questions: How do we govern it? How do we validate it? How do we make it economically viable? How do we use it responsibly on real matters?

The organisations making the most progress are not necessarily the ones adopting the most tools. They are the ones building repeatable processes, establishing clear guardrails, and approaching AI as an enhancement to legal expertise rather than a replacement for it.

As legal AI continues to evolve, those fundamentals will matter far more than any individual platform or model. And if the candid discussion at NAMWOLF was any indication, the legal industry is increasingly ready to move from experimentation to responsible implementation.

For more regarding Cimplifi specialised expertise regarding AI & machine learning, click here.

About the Author
Sashi Valavala leads the development of defensible AI and analytics solutions across the legal data lifecycle. With over 18 years of experience in eDiscovery analytics and consulting, he designs sound workflows for matters, including HSR Second Requests and complex litigation. Sashi brings a collaborative, practical approach to AI adoption and results.

 

 

 

Schedule a live, personalised demo with Cimplifi today or click here to learn more about Maestro.