Editor’s Note: Generative AI is no longer a pilot project in eDiscovery—it’s now a production asset. Drawing from the final results of the 2H 2025 eDiscovery Business Confidence Survey, this technology-focused installment in ComplexDiscovery’s ongoing series highlights how the industry has decisively moved from testing to deployment of Large Language Models (LLMs) and Generative AI (GAI). With insights from 64 industry professionals—primarily from law firms and software/service providers—the findings reveal that improved quality, not just cost savings, is driving adoption.

Yet significant concerns remain around trust, accuracy, and defensibility. For cybersecurity, information governance, and legal operations leaders, this article surfaces both the opportunities and urgent responsibilities that come with this rapid AI integration. It is the second in a four-part series analyzing Market Sentiment, Technology Integration, Operational Metrics, and Strategic Challenges in the eDiscovery ecosystem.


Content Assessment: The Shift from AI Pilots to Production: Insights from the 2H 2025 eDiscovery Business Confidence Survey

Information - 94%
Insight - 93%
Relevance - 93%
Objectivity - 94%
Authority - 95%

94%

Excellent

A short percentage-based assessment of the qualitative benefit expressed as a percentage of positive reception of the recent article from ComplexDiscovery OÜ titled, "The Shift from AI Pilots to Production: Insights from the 2H 2025 eDiscovery Business Confidence Survey."


Industry Research

The Shift from AI Pilots to Production: Insights from the 2H 2025 eDiscovery Business Confidence Survey

ComplexDiscovery Staff

If 2024 was the year of “AI exploration” in legal technology, the second half of 2025 has unequivocally become the era of “AI execution.” The results of the 2H 2025 eDiscovery Business Confidence Survey reveal a market that has moved past the initial hype cycle and is now deeply entrenched in the practical realities of deployment. For eDiscovery professionals, the debate over whether to adopt Large Language Models (LLMs) and Generative AI (GAI) is effectively over; the conversation has now shifted entirely to how to manage the risks and realities of a transformed workflow.

The Tipping Point: Active Deployment Dominates

The most striking statistic from the 2H 2025 survey is the sheer scale of adoption. A commanding 64.06% of respondents report that their organizations are currently “Integrating and Deploying” LLMs and GAI technologies. These tools are no longer theoretical; they are either already deployed or in the process of being integrated into operations and offerings. When combined with those “Testing and Piloting” (10.94%) and those “Considering and Evaluating” (18.75%), nearly the entire industry is engaged with this technology. Perhaps most telling is the shrinking minority of holdouts: only 6.25% of respondents indicated they have “No Plans” to explore or integrate these technologies. This holdout effectively signals that non-adoption has become a competitive disadvantage rather than a conservative safety play.


Use of LLMs and GAI in Organization's Operations or Offerings - 2H25

The Primary Driver: Quality Over Cost

Contrary to the popular narrative that AI is primarily a tool for slashing budgets, the survey data suggests a different motivation among eDiscovery leaders. When asked for the primary benefit of integrating these tools, the majority (54.69%) cited “Improved Service/Product Delivery.” Respondents are looking to AI to enhance the speed, quality, and range of their deliverables. In contrast, “Cost Savings” was cited as the primary benefit by only 12.50% of respondents. This result indicates that the industry views Generative AI as a value-add technology—something that improves legal work and speeds it up—rather than merely a mechanism for commoditization or headcount reduction.


Primary Benefit of Integrating LLMs and GAI into Organization's Operations or Offerings - 2H25

The Trust Gap: Accuracy Remains the Achilles’ Heel

However, the road to full integration is paved with skepticism. Despite the aggressive deployment rates, the “Trust Gap” remains the single largest barrier to success. When asked about the primary challenge of integrating LLMs, nearly one-third of respondents (32.81%) pointed to “Results Accuracy,” specifically citing concerns about hallucinations and data reliability. This fear outweighs financial concerns; “High Costs” (18.75%) and “Unclear ROI” (17.19%) are significant but secondary hurdles. The industry is willing to pay for these tools, but they are terrified of being wrong. This anxiety is well-founded in a profession where a single “hallucination” in a court filing can result in sanctions or reputational ruin.


Primary Challenge of Integrating LLMs and GAI into Organization's Operations or Offerings - 2H25

Implications for the Wider Ecosystem

For eDiscovery professionals, these findings dictate the immediate strategic roadmap. With 64% of peers already deploying, the pressure is on to establish defensibility protocols and move from enthusiasm to evidence. The central accuracy concern suggests that the next competitive battleground will not be who has the most capable AI, but who has the most verifiable AI, supported by documented workflows and human-in-the-loop validation.

For Cybersecurity professionals, the survey exposes a glaring blind spot. Only 1.56% of respondents viewed “Risk Mitigation” as the primary benefit of LLMs. This response suggests that while the industry is rushing to use AI for output (drafting, summarizing), it is largely ignoring AI as a tool for defense (anomaly detection, threat hunting). Furthermore, the rapid deployment of these models introduces significant new attack vectors, from prompt injection to training data poisoning. Security leaders must urgently audit these “Integrating and Deploying” environments to ensure that the drive for “Improved Service Delivery” hasn’t opened backdoors into sensitive client data.

For Information Governance (IG) professionals, that same accuracy challenge becomes a data quality mandate. If the information feeding these models is “ROT” (Redundant, Obsolete, Trivial), hallucinations are not an anomaly but an inevitability. IG leaders must position themselves as gatekeepers of data provenance and lifecycle, ensuring the fuel for these systems is clean enough to minimize the very risks that survey respondents fear most.

Ultimately, the 2H 2025 survey confirms that Generative AI has graduated from the lab to the front lines of eDiscovery. The widespread adoption is driven by a desire for quality and speed, but it is checked by a persistent fear of inaccuracy. As the industry matures, success will belong to those who can bridge the gap between the potential of the technology and the precision required by the law.

Coming Next in the Series: We know the industry is adopting AI, but is the financial house in order? In Part 3, we uncover the “Visibility Gap” in operational metrics, examining the hidden risks in Days Sales Outstanding (DSO) and Monthly Recurring Revenue (MRR).



News Source
 
Robinson, R., & Robinson, H. (2025). 2H 2025 eDiscovery Business Confidence Survey by ComplexDiscovery OÜ and EDRM. ComplexDiscovery OÜ.

Assisted by GAI and LLM Technologies

Additional Reading

Source: ComplexDiscovery OÜ

 

Have a Request?

If you have information or offering requests that you would like to ask us about, please let us know, and we will make our response to you a priority.

ComplexDiscovery OÜ is a highly recognized digital publication focused on providing detailed insights into the fields of cybersecurity, information governance, and eDiscovery. Based in Estonia, a hub for digital innovation, ComplexDiscovery OÜ upholds rigorous standards in journalistic integrity, delivering nuanced analyses of global trends, technology advancements, and the eDiscovery sector. The publication expertly connects intricate legal technology issues with the broader narrative of international business and current events, offering its readership invaluable insights for informed decision-making.

For the latest in law, technology, and business, visit ComplexDiscovery.com.

 

Generative Artificial Intelligence and Large Language Model Use

ComplexDiscovery OÜ recognizes the value of GAI and LLM tools in streamlining content creation processes and enhancing the overall quality of its research, writing, and editing efforts. To this end, ComplexDiscovery OÜ regularly employs GAI tools, including ChatGPT, Claude, Grammarly, Midjourney, and Perplexity, to assist, augment, and accelerate the development and publication of both new and revised content in posts and pages published (initiated in late 2022).

ComplexDiscovery also provides a ChatGPT-powered AI article assistant for its users. This feature leverages LLM capabilities to generate relevant and valuable insights related to specific page and post content published on ComplexDiscovery.com. By offering this AI-driven service, ComplexDiscovery OÜ aims to create a more interactive and engaging experience for its users, while highlighting the importance of responsible and ethical use of GAI and LLM technologies.