Insights Articles

• David Horrigan (Moderator) - Discovery Counsel & Legal Education Director at Relativity
• Julia Helmer - Director, Client Solutions at Array
• Matt Smith - EVP, Business Development at Array
• Maks Babuder - Director of Product at Relativity
• Cristin Traylor - Senior Director, AI Transformation & Law Firm Strategy at Relativity

Julia Helmer observed that teams are moving beyond the idea of choosing between generative AI and machine learning. Instead, they’re designing hybrid workflows that leverage both. As she explained, generative AI excels at surfacing early insights—summaries, topics, contextual rationales—while continuous active learning (CAL) remains essential for precision, stability, and defending outcomes under scrutiny. “The most effective workflows,” she noted, “are the ones that pair the early visibility of GenAI with the defensibility of CAL.”

Relativity’s Maks Babuder reinforced this point with data drawn from two years of aiR deployments. More than 250 customers have now adopted aiR, generating over 200 million document predictions, with more than twenty organizations each processing upward of a million documents. Across these matters, Maks has consistently seen linear review speeds improve by more than 50 percent, as reviewers use aiR’s summaries, predictions, and explanations to accelerate their understanding of each document. Beyond traditional review, aiR is now powering use cases like internal investigations, DSARs, FOIA requests, breach response, and incoming production analysis—evidence that AI is becoming a layer woven throughout the entire discovery lifecycle.

Adding another dimension, Cristin Traylor highlighted how this shift is elevating workflows from simple responsiveness decisions to deeper case intelligence. She noted that AI is helping teams identify key actors, map timelines, and pinpoint critical facts far earlier than before, enabling lawyers to advise clients faster and with greater context. This movement away from document‑centric thinking toward fact‑centric insight represents a meaningful change in how matters are understood and strategized.

Across the panel, a recurring theme emerged: technology alone doesn’t drive transformation—organizational behavior does. Cristin explained that successful adoption depends on both top‑down direction and bottom‑up enthusiasm. Leadership must signal that AI belongs inside workflows, but real progress depends on “people who are hungry to implement,” she said. Without that dual force, even the most powerful tools stall.

Julia and Matt echoed this sentiment, emphasizing that high‑performing teams tend to share several traits: they start small, choose workflows that benefit immediately from AI, and then scale what proves effective. They identify internal champions who help normalize new behaviors. They anchor their approach in workflow design, not feature chasing. And they recognize that incremental progress is often more sustainable than wholesale reinvention.

Matt also noted that generative AI feels dramatically more accessible than earlier AI waves, especially when compared to TAR. While TAR required familiarity with specialized workflows, generative AI benefits from the fact that “everyone is using it in some way in their daily lives,” he said, which collapses the learning curve and lowers the emotional barrier to adoption.

Maks emphasized that the real expertise emerging inside organizations isn’t tied to prompt engineering but to understanding how workflows behave, how the technology behaves, and how to guide matter teams in aligning the two. In his words, “Actionable insights only matter if users trust them. And trust comes from understanding the process.”

The panel also explored how shifting market expectations are influencing legal teams' AI strategies. Matt noted that RelativityOne’s inclusion of aiR has created more predictable pricing models, reducing hesitation for organizations that once worried about whether they would “get enough juice out of the squeeze.” With greater transparency into cost and capability, teams are more confident experimenting and scaling their use of AI.

Julia addressed the concern many clients initially bring to conversations—hallucinations—and shared how quickly these concerns fade once clients see aiR’s grounded citations, rationales, and the system’s ability to flag statements it cannot substantiate. The transparency of the output makes the technology inherently easier to trust, she said, noting that “as soon as clients see how the system explains itself, the worry falls away.”

Cristin added that the shifting expectations are not just about tools but about roles. Junior attorneys and technologists are increasingly contributing earlier to case strategy, thanks to AI absorbing so much of the manual review burden. She put it simply: “If you can draft a review protocol, you can draft a prompt”—a reminder that the skills required to use AI align naturally with existing legal competencies.

Looking forward, Maks offered a preview of the enhancements coming to Relativity aiR in 2026. These include aiR Assist, which will bring natural‑language search to early case assessment; new analysis types for Confidential Business Information and Personal Information with auto‑redaction; significant upgrades to privilege workflows; and custom analyses that extend aiR's intelligence into images, PDFs, and other non-traditional formats. He emphasized that these innovations “aren’t hypothetical—they’re grounded in how teams actually work,” and they reflect the growing expectation that AI should be embedded throughout the lifecycle, not treated as an isolated tool.

Cristin added that these innovations are particularly poised to transform case strategy work, noting that tools like aiR for Case Strategy are enabling teams to extract facts, identify missing pieces, and create deposition materials far earlier than the traditional review‑first model ever allowed.

Across every discussion point, one message resonated clearly: the organizations thriving with AI are not the ones chasing features—they’re the ones designing workflows.

Competitive advantage in 2026 will come from teams that blend generative AI with defensible review methods, standardize and validate their processes, strengthen internal comfort with explainability, and embrace hybrid roles that combine legal and technical fluency. AI is no longer reshaping only how documents are reviewed—it’s reshaping how matters are understood, staffed, budgeted, and strategized.

The question moving forward is not whether legal teams will use AI, but how intentionally, intelligently, and operationally they bring it into their practice.