The Future of eDiscovery: Generative AI’s Impact – Jon Fowler, Secretariat

In this interview, we had the chance to sit down with Jon Fowler, Managing Director of Secretariat, a sponsor of Future Lawyer UK. Jon shares his insights on how generative AI is reshaping eDiscovery, the evolving role of technology in litigation, and what legal teams need to consider as they navigate this rapidly changing landscape.

Enjoy the blog, and get ready for the conversation next week on Litigation Day, where we'll explore the rapidly evolving landscape of disputes and eDiscovery.

Secretariat has observed a significant shift in the legal industry with the rapid adoption of generative AI, particularly among litigation-focused law firms. Could you share how you have seen clients using/wanting to use generative AI as part of the eDiscovery process, and how do you see this developing in the short to medium term?

There has been significant discussion around generative AI (GenAI) in the legal industry, but in practice, its adoption (particularly in eDiscovery) is still at a fairly early stage. While clients are eager to understand how others are leveraging this technology, the actual implementation has not yet caught up with the level of interest and industry buzz.

One implementation generating considerable attention is Relativity aiR, which focuses on integrating generative AI into first-pass document review. The technology works by applying document review protocols to predict whether a document is responsive or privileged. These predictions are accompanied by both a recommendation and an analysis of the opposing rationale—explaining why the document might not be responsive or privileged. Last year, Relativity piloted aiR with its customers, and, particularly in the United States where uptake has been greater, law firms have reported positive results from testing.

However, there is still some debate about how GenAI may alter Technology Assisted Review (TAR) protocols and how it will be incorporated into electronically stored information (ESI) agreements with courts. Many firms remain hesitant to fully adopt this approach due to cost concerns and the lack of judicial precedents that would provide confidence in its acceptance. As the technology matures, costs decline, and case law begins to recognise GenAI-assisted processes, we anticipate a tipping point where adoption will accelerate.

In the meantime, clients are showing greater comfort with generative AI applications that offer lower risk and cost, such as summarisation and advanced search functionalities. Many platforms now enable users to interact with GenAI-driven tools to query datasets, summarise documents, and surface key facts more efficiently. These capabilities are proving to be a practical entry point for integrating generative AI into the eDiscovery process.

To gain further insight into how AI is being used in legal workflows, Secretariat is releasing an AI survey during Legal Week in New York. One key finding of the survey is that 77% of respondents have used GenAI technology in their work over the last 12 months. The most frequently used applications include ChatGPT and Microsoft Copilot; with the primary use cases being legal research, drafting documents, document review, and summarising content and data.

With your industry-leading expertise in digital forensics and eDiscovery, how has Secretariat leveraged AI and machine learning to enhance data analysis and streamline investigative processes?

AI and machine learning have become integral to our approach, playing a significant role in most matters we handle. Whilst many of these technologies have been part of the eDiscovery industry for over a decade, their application varies depending on the specific needs of each matter. Rather than applying every solution indiscriminately, we take a strategic approach identifying the key objectives and leveraging the right tool for the job

For instance, we employ predictive coding when data volumes justify its use, apply visual analytics and clustering for pattern recognition, utilise machine learning for translations, and deploy generative AI technologies when valuable to the use case. We have extensive experience with industry leading technologies including Relativity’s Active Learning and its aiR for Review and aiR for Privilege products, which help streamline document review workflows and enhance efficiency.  

Additionally, we have developed proprietary machine learning models designed for the prediction of document characteristics, which operate outside of traditional document review platforms. These predictions are then fed into the review platform allowing for a more refined and targeted review process. This ability to build and customise predictive models enables us to provide more flexible and innovative AI-driven solutions tailored to each case. 

At Secretariat, our approach is always case-by-case. Every matter is different, so we assess the data, the objectives, and then decide what tools make the most sense. The goal is always the same: efficiency, accuracy, and getting to the insights that truly matter as quickly as possible.

Secretariat's Data Solutions practice expands the firm's capabilities in data-driven disputes and investigations. Can you discuss the impetus behind creating this practice and how it enhances your ability to serve clients in complex legal matters?

Secretariat’s origins are rooted in expert advisory, initially focusing on construction expertise, arbitration, and damages. However, data has become a critical component across all these areas, and we quickly recognised the need for a dedicated practice to address the growing complexities of data-driven disputes and investigations.

Since establishing the Data Solutions practice, we’ve seen significant synergies across different practice areas. By integrating data expertise with our broader advisory services, we can deliver more intelligent, comprehensive solutions for our clients. This practice is also closely aligned with our global Investigations & Disputes team, which is Secretariat’s fastest-growing area. Given the intrinsic link between investigations and data analytics, having a robust data capability is essential to providing clients with deeper insights and more effective outcomes.

Additionally, the Data Solutions practice enables us to develop and deploy innovative data-driven products tailored to our clients' evolving needs. While it is still a relatively new offering, it has quickly become a vital part of our broader service capabilities, enhancing our ability to navigate complex legal matters with precision and efficiency.

Having a dedicated team focused on data solutions means we can provide a more seamless service to our clients, whether it’s eDiscovery, structured data analysis, or AI-driven analytics. Ultimately, it’s about delivering smarter solutions. By integrating data expertise across different parts of the business, we’re helping clients get more value from the information they already have, rather than treating data analysis as a separate function. That’s where the real advantage lies.

In discussions about generative AI's impact on eDiscovery, it's noted that while upfront technology costs have increased, reliance on manpower has drastically decreased, leading to lower overall expenditures. How is Secretariat adapting its eDiscovery services in response to these changes, and what cost benefits will clients realise?

That’s an interesting point, and it largely ties back to the generative AI-driven document review process I mentioned earlier. While technology costs are currently high—given the per-document and per-data processing fees associated with first-level review—the long-term financial impact is expected to be positive. Once confidence in these AI-driven solutions reaches a critical level, they will significantly reduce the need for lawyers and paralegals to conduct first-pass document reviews. Given the scale of large eDiscovery matters, where millions of documents may need review, the cost savings from reducing billable legal hours will likely outweigh the increased technology expenditure.

However, we’re not quite at that tipping point yet. While generative AI for first-level review is likely to become a standard approach in the future, for now, legal teams are still assessing its reliability. When adoption does become widespread, it will allow lawyers to focus their expertise on higher-value tasks, such as second-level reviews and more substantive legal analysis. This evolution is similar to what we saw over a decade ago with the introduction of predictive coding. Initial fears that it would replace lawyers never materialised; rather, it shifted legal work to areas where human expertise adds the most value.

In terms of how Secretariat is adapting, our approach remains consistent. We do not provide managed review services, so these changes do not directly impact our service model. Instead, our focus has always been on helping clients maximise efficiency and cost-effectiveness. Whether through traditional review processes or AI-driven solutions, we continue to advise clients on the most strategic and financially sound approaches to eDiscovery.

For larger providers with offshore review centres, these shifts will likely drive significant operational changes. If I were managing a large-scale offshore review team, I would be carefully considering how to optimise the use of resources over the next five years. But as an expert consulting firm, Secretariat’s role remains unchanged—helping clients navigate these advancements to achieve the best possible outcomes.

What key insights or solutions will Secretariat bring to Future Lawyer UK 2025, particularly regarding the integration of AI and data-driven technologies in the legal sector?

The key insight Secretariat brings to Future Lawyer 2025 is expertise. Sharing real, hands-on experience in how AI is actually being used in legal workflows and highlighting the critical need for independent expertise in navigating the rapidly evolving AI landscape. It’s not just about adopting new technology; it’s about integrating it properly to get real value. 

AI is reshaping legal practice, from eDiscovery to predictive analytics, and while the industry has been quick to explore its potential, the real challenge lies in effectively integrating these technologies into legal workflows. Adoption isn’t just about accessing new tools—it’s about understanding when and how to implement them to maximise efficiency and strategic value.

As I mentioned earlier, a major theme that we are seeing is the tipping point between cost and acceptance. AI-driven solutions, including those for document review, predictive analytics, and risk assessment, are advancing rapidly. However, widespread adoption depends on two key factors: confidence in the technology’s reliability and a clear return on investment. Legal teams are already leveraging AI to assess case outcomes, manage risks, and refine settlement strategies by analysing historical data and judicial patterns. These tools can provide valuable insights into judicial behaviour and opposing counsel tendencies fundamentally changing how cases are approached. Yet, as with eDiscovery, firms are still weighing the trade-offs between human expertise and AI-driven efficiency.

At Secretariat, our role is to help clients navigate this shifting landscape, ensuring they implement AI solutions in a way that drives real value rather than just following industry trends. With decades of experience at the intersection of legal technology, AI, and analytics, we provide independent expertise that enables firms to make informed, strategic decisions about AI adoption. As the industry approaches this tipping point between cost and acceptance, those who proactively invest in expertise and thoughtful implementation are the ones who will see the biggest benefits.

---

Meet the Secretariat team and Jon on Litigation Day, 3rd March 2025. Jon will join Panel 3, sharing his insights on “AI-Augmented Legal Strategy: Transforming Litigation Practice Through the Legal and End-Client Lens.”

Reserve your spot now and join the conversation next week!