The legal industry has long grappled with an ever-growing volume of data. As matters grow in size and complexity, the tools legal teams rely on to process that data must evolve just as quickly. At Array, we help clients navigate increasingly demanding workflows with speed, precision, and scale, combining experienced professionals with technology that helps them manage evolving data challenges.
Relativity recently sat down with Julia Helmer, JD, vice president of advisory and analytics at Array, and Eric Wendt, senior product manager at Relativity, to explore the role of processing in modern eDiscovery. The conversation examines how processing capabilities have evolved over the past decade, the data challenges facing legal teams today, and what's ahead for practitioners looking to move from collection to review faster and smarter.
Questions:
When you compare modern day projects to what you were managing five or ten years ago, what feels most different, and where do you see challenges show up now?
Julia: Of course, sheer volume is a huge difference. Ten years ago, a large matter was maybe 50,000 documents. That is now a small matter, and most mid-sized matters have several million documents with our largest at over 70 million. We’re also seeing a huge jump in mobile and short message data; it’s become a standard in almost every matter we have to engage collections for. Certainly since COVID, the amount of short message data sources has dramatically increased.
If we were to think about how teams approached processing in the past compared to how they approach it now, what’s would you say is the biggest shift?
Eric: I think the largest shift has been from understanding how to process large amounts of data at scale to how we can extract insights and learn from it. It’s no longer a question of, “Can we process these 20 million documents?” but of, “what quick insights can we pull from these 20 million documents?” We’ve moved beyond many operational complexities like converting data types, managing job sizes, or managing priority queues and are now asking more thought-provoking questions about the data. This is where workflows like Repository workspaces and Early Case Intelligence complement initial processing. Now that processing has become a simpler task to scale, we can focus on more high-value project work.
Can you share a real-world example of your experience working with Processing in Relativity aiR and how it has helped Array manage complex data challenges with your customers?
Julia: Processing in Relativity aiR has allowed us to run multiple data sources through the engine without needing to convert those sources ahead of time, speeding up the time between collection and data access for our clients. As an example, we used to have to take UFDR data from cell phone collections and send them to a third party or use a separate software to convert those exports to Relativity Short Message Format – often losing a day or more in the pipeline. Now that we can directly process UFDR data with Relativity aiR, we are able to have data available to our clients more quickly and more securely than ever before.
Eric: Processing’s ability to scale its compute resources truly sets it apart from other tools. We’ve committed a lot of energy over the past several years to take advantage of the latest cloud-native technologies, powering the most complex legal data intelligence workflows. This effort resulted in tripling our throughput of processing speeds and we plan on increasing that further in 2026. In addition to raw performance, using tools like Early Case Intelligence will give customers fast insights into ever growing data volumes.
Understanding your data makeup before conducting a more specific analysis is going to be a gamechanger for processing.
Julia: Some challenges we have had include processing of file types that have not traditionally lent themselves well to review platforms, such as CAD files or multi-layered PDFs. Another challenge is the ability to process multiple collections for the same project simultaneously – some processing throughput is handled serially rather than concurrently, which has inhibited multitasking. Modern attachments have presented some new, unique challenges, too, with matching up linked documents to the emails in which they are referenced.
Eric: We're committed to serving the most modern data types, so we're always keeping a pulse on industry shifts. It's a fun challenge, as there is always something new on the horizon.
One of the clearest signals we've seen recently is what's happened with short message data — specifically with RSMF trends. Two years ago, the overwhelming majority of RSMF conversions were being handled by third-party tools. The fact that that's shifted significantly, with Processing in Relativity aiR now handling the majority of that volume, tells us something important: when practitioners have a capable native option in the platform they're already using, they prefer it. Conversion that used to mean an extra step or an extra day in the pipeline now happens where everything else happens. Beyond speed, customers tell us that consolidating conversion inside Relativity reduces risk: fewer handoffs, fewer tools in the chain of custody, and one defensible workflow.
We're seeing a similar dynamic play out with audio and video. These file types have historically sat outside the standard review workflow — teams either handed them off to vendors for manual transcription or simply deprioritized them. But A/V data is showing up in more matters and more consistently, and the expectation that it should be searchable and reviewable alongside everything else is growing. The ability to bring that evidence into the same workflow, rather than treating it as a special-case exception, is becoming less of a nice-to-have and more of a baseline requirement.
That's informed a lot of where we've been investing. Lately, we've focused on adding support for things like Apple iWorks file types and native Google Chat conversion to RSMF. We've also been addressing the challenge of linked cloud attachments out of Google Vault exports, with Microsoft Office 365 links next on the roadmap. And we're not only focused on what's new — we'll also be improving our support of EnCase images.
Julia: Speed and efficiency are probably the most critical portion of the document processing workflow. We want to minimize as much “machine time” as possible because our case teams are always chomping at the bit to have access to their documents so they can begin their review. Particularly in unique situations like regulator-initiated second requests, the clock is constantly ticking in the background – so the sooner we can release the documents for review, the more time our teams have to analyze the data. Teams can leverage people, process, and technology to create a solid, repeatable plan that optimizes for efficiency. Using Relativity aiR, ensuring that our team is creating the most efficient and defensible workflows available, and then monitoring those processes to identify areas of improvement can help accelerate these projects individually and over time.
Eric: I think the most exciting developments come from our increased usage of AI. Processing in Relativity aiR is not only there to provide a foundational, defensible data layer, but also to pull insights from the content it's ingesting.
Automatically extracting insights like collection coverage, document themes, or finding some interesting communications between key individuals earlier in the case lifecycle allows project teams to make more informed decisions earlier than ever. It also frees precious time formerly spent manually extracting that data on other, high-value work. More time can be spent focusing on answering questions pertaining to the case itself, rather than sorting out all the processed content manually.
Can you discuss specific features of Processing in Relativity aiR that have transformed your internal workflows, and how have they impacted case outcomes?
Julia: Relativity’s ability to process Google Workspace data has had a large impact on our workflows. Array has several clients that rely on the Google Suite and Google Vault to manage their data. In the past, the approach for these data types has been to treat them similarly to on-premises email and data storage systems, but since they’re not precisely the same, that can lead to the loss of important data that our case teams may need to prove their case. For example, having access to previous versions of documents, seeing when a document was changed in Google Drive, or knowing who had access to a specific document may be important pieces of the story that lawyers are trying to tell.
Relativity aiR’s ability to read and map data from the Google Metadata Sidecar file is so critical for our clients to be able to capture and access this kind of key information they may need to prove up their case. This ability has impacted case outcomes because clients are able to quickly see and analyze that information in one system of record, and take action on it, without needing to wade through a ton of external exports, charts, and files to try to reconstruct what happened with a specific document.