Beyond the Buzz: How Legal Teams Are Really Using AI

Recap from the Masters Conference Denver | Right Discovery Staff Writer

At the 2025 Masters Conference in Denver, one breakout session stood out not for lofty predictions, but for its real-world application of AI in legal practice—and the candid truths that came with it. “AI Stories From the Front Line: Real Life Stories of Using AI in eDiscovery and Legal” - the session pulled back the curtain on how law firms and legal departments are already deploying AI in privilege logs, relevance review, and even inbox management. This wasn’t about hype—it was about how artificial intelligence is quietly, and sometimes radically, transforming the day-to-day work of legal professionals.

Leading the conversation were Sloane Dryer (Holland & Hart) and Pete Feinberg (Consilio), who brought the audience behind the scenes of their own legal teams’ experiences—showing how AI is moving from theory to trusted tool in day-to-day legal practice. And it is important to note that “If you’re not using AI yet, you’re already behind.”

One of the most compelling use cases discussed was the application of generative AI in drafting privilege log entries. In one high-volume matter involving over 89,000 documents, a legal team deployed AI to generate initial privilege descriptions. These drafts were reviewed by attorneys for accuracy and quality control. Impressively, over 50% of the AI-generated entries were accepted without modification—demonstrating the growing reliability of AI-assisted review in large-scale litigation.

In the realm of document relevance determinations, the panel cited a case study in which generative AI outperformed human reviewers in accuracy. According to the speaker, the AI system achieved 96% accuracy and included explanatory rationale and contextual excerpts to support its conclusions—providing not just results, but traceable reasoning. “The machine had backup,” one presenter remarked, “and in most cases, it was right.”

Email management was another area where AI is quietly making a significant impact. Tools like SimplyFile, an Outlook extension that learns user behavior to automatically suggest email filing folders, offer a simple but effective entry point into AI-driven productivity. While not a headline-grabbing technology, it reflects the growing trend of embedding AI into everyday legal operations.

Standard tools like Zoom’s AI often produce overly generic outputs. To address this, Sloane Dryer described a proprietary system built on foundational models like Claude and OpenAI. This internal process involves downloading transcripts, converting them to Word format, and running them through custom prompt structures to extract summaries, identify key action items, and maintain privacy compliance. Feedback from internal users indicated that the quality of these AI-generated summaries far exceeded expectations.

The session also explored the boundaries of AI capabilities. Current models still struggle with structured data formats—such as spreadsheets—and are not yet dependable for final privilege determinations. The risk of AI “hallucinations” remains, which reinforces the need for rigorous human oversight. To mitigate these risks, the presenters emphasized careful prompt engineering, encouraging practitioners to include instructions that allow for uncertainty, require cited sources, and guide the AI to act as an assistant—not a decision-maker.

Throughout the discussion, both Sloane and Pete repeatedly highlighted the importance of context-rich prompting, thoughtful constraint setting, and clear delineation of audience and tone. These practices, they explained, are key to minimizing inaccuracies and maximizing the usefulness of AI-generated outputs.

Key Takeaways

• Generative AI significantly accelerates privilege log drafting, with over half of entries accepted unedited in a recent 89,000-document review.

• AI outperformed human reviewers in relevance coding, achieving 96% accuracy and offering supporting rationale.

• SimplyFile for Outlook offers practical, low-cost automation for email organization.

• Zoom AI summaries are limited, but internal tools leveraging Claude/OpenAI provide high-quality transcript summarization.

• Prompt design is critical—clarifying audience, tone, and expected structure dramatically improves AI output reliability.

• AI should supplement human work, not replace legal judgment—especially in complex or privileged matters.

• Structured data remains a challenge for current AI models; spreadsheets and chat messages require human review.

• Human review and error handling remain essential components of any responsible AI workflow.

“AI isn’t just changing how we practice law—it’s changing what’s possible.” This session wasn’t just a conversation about technology—it was a challenge to the industry to think bigger, move faster, and adapt smarter. AI won’t replace good lawyers, but it will absolutely expose inefficient ones. The message from Denver was clear: the firms that lean in now—strategically, responsibly, and with a healthy dose of skepticism—will shape the future of legal practice. The rest? They’ll be catching up.

AI may be accelerating legal workflows, but it also introduces new complexities in compliance, risk, and data management. Having the right partners makes all the difference. Stay ahead of what’s next. Follow Right Discovery for updates on legal tech, AI innovation, and cyber resilience—or contact us at solutions@rightdiscovery.com to learn how we can support you.

A special thanks to our panelists for their time and for sharing their experiences and insights:

Sloane Dryer, Holland & Hart

Pete Feinberg, Consilio

Right Discovery is a proud sponsor of the Masters Conference & Masters Conference Legal Events.

Need help getting AI off your “someday” list? Contact Right Discovery at solutions@rightdiscovery.com—we’ve got you covered. Follow us for the latest insights and join us at the year’s most anticipated legal events.

Topics: generative AI, legal document review, email organization, SimplyFile, Claude, OpenAI, AI prompt engineering, Zoom AI, structured data, meeting summarization, AI-powered email tools, legal technology, Masters Conference, document coding, AI workflows, legal innovation, privacy compliance, AI risk management, legal event