If you have ever tried to explain artificial intelligence to a room full of legal professionals, you know the challenge. Half the audience leans forward in curiosity, and the other half folds their arms in caution. The 2025 Sedona Conference Working Group 1 Annual Meeting in Charlotte captured that exact mix of energy, pragmatic skepticism, and genuine progress.

The Sedona Conference Working Group 1 is one of those rare places where eDiscovery professionals, litigators, and judges set aside the usual competition to focus on collaboration and finding compromise. It is where litigators and the legal technology industry pause to ask, “How do we move forward together?” This year, the conversation about AI was far from just a hypothetical exercise. It was practical, grounded, and refreshingly candid.

Cautious Optimism: When Curiosity Meets Responsibility

Across sessions, there was a strong sense of optimism about AI’s potential to improve eDiscovery and legal workflows. Many participants described a climate of “cautious optimism,” and I think that reflects where many are today.

Some organizations are moving quickly and seeing measurable results. Others are taking a slower approach, starting in lower-risk areas and learning through pilot projects. Those who have been experimenting and refining their approaches for several years seem to have the most confidence in their deployment of AI-based solutions and workflows.

Verification and defensibility were also central themes throughout the meetings. No one claimed to have a single accepted formula for validating AI-driven results. Instead, the focus was on developing processes that would instill confidence through as much transparency as necessary to build trust with opponents and the courts. That included clarity about what needs to be validated, how those validation methodologies should be performed, and how much documentation is appropriate (and possible).

Additionally, the judges in attendance offered a reassuring perspective. Many expressed a willingness to give parties a measure of grace when they use AI thoughtfully and in good faith. It shows that the courts recognize the potential benefits of these technologies while still expecting accountability. But, as with any group of practitioners, the perspectives offered by the judges were far from uniform. It reflects the broader reality that even the most tech-savvy judges still have limited time to really understand how emerging legal technologies work, including what is and is not possible.

Trust Through Transparency

Transparency was another strong theme. Attendees debated how much information should be shared about the use of AI tools. Some felt reluctant to share their work product, but others agreed that being transparent about what tools are used, how they are integrated into the process, and how the results are verified will go a long way toward building trust. The consensus seemed to be less about exposing sensitive details and more about demonstrating that the process behind the technology is reliable and fair.

Adoption levels varied, but it was clear that AI usage is no longer a hypothetical discussion. A large portion of attendees indicated that they are already using AI in areas like deposition preparation, data analysis, and early case assessment. Adoption is slower for higher-risk tasks, such as privilege and responsiveness review, although that gap appears to be narrowing as confidence grows.

From my perspective, courts and practitioners are still working toward alignment on what “reasonable” use of AI looks like. The challenge is that we are not all referring to one type of technology or one workflow. The term “AI” encompasses a wide range of tools, each addressing different problems. Achieving consistency across such diversity is difficult. With such variance, the most effective path forward starts with identifying clear objectives for each process, designing workflows, and choosing technologies that meet those objectives.

Process-Driven AI: Structure Before Technology

If there was one lasting takeaway from the Sedona meeting, it is that AI succeeds when the underlying process is strong. In my experience, most frustration with technology does not come from the tools themselves. It comes from the lack of thoughtful and tested workflows surrounding them. When teams invest time in defining their goals, roles, and checkpoints, the outcomes are more predictable, easier to evaluate, and easier to defend.

For organizations preparing to implement or expand their use of emerging technologies, the best first step is self-assessment. Identify what problems need solving and what improvements matter most. Whether the goal is efficiency, cost control, or quality, starting small and expanding based on results tends to build lasting confidence. Sedona revealed that the most successful teams learn by doing, and they do it in controlled, intentional ways.

At TCDI, we take that same approach with our clients. We begin by listening carefully and understanding their processes and priorities before recommending specific tools. We also stay proactive in evaluating new technologies so that when a client presents a challenge, we already have insight into what solutions might fit best. Through our LitForward Center for Technology, Research and Analysis, we are constantly exploring new and developing solutions in a trusted setting. Being part of collaborative forums such as the Sedona Conference Working Groups furthers this mission as we continue to refine emerging ideas to best address our clients’ most complex data challenges.

Conclusion: Reflections from Sedona

Finally, I want to express appreciation for The Sedona Conference and the important role it plays in this evolving landscape. Its balanced approach, mix of voices, and commitment to thoughtful progress is invaluable. The work that happens there continues to move the law forward in a just and reasoned way. I also want to thank my colleague, Tim Opsitnick, for his service on The Sedona Conference Board of Directors. Without passionate judges and litigators like Tim, who work tirelessly to progress the legal industry, valuable gatherings like Sedona WG1’s Annual Meeting would not be possible.

I left Charlotte encouraged by the openness of the discussions and the shared commitment to learning. The spirit of collaboration that defines Sedona will help all of us continue to build trust, refine processes, and move the practice of eDiscovery forward responsibly.

Andy Cosgrove

Author

Share article:

Andy Cosgrove is the Chief Strategy Officer of TCDI. He is responsible for guiding the development of innovative technologies and strategies that enable TCDI clients to mitigate legal risks and minimize costs in the areas of eDiscovery, information governance, cybersecurity and data privacy.

Andy brings over 20 years of experience as both outside counsel at AmLaw 100 law firms and, most recently, as in-house managing counsel leading information governance and eDiscovery teams at Juul Labs and General Motors. Learn more about Andy.