When a Second Request arrives, one of the first questions legal teams ask is, "What data do we need to collect?"
A better question is often, "What data do we actually have?"
The difference matters.
When responding to a Second Request, organizations are often required to identify, preserve, collect, review, and produce information from across multiple systems, custodians, and business functions. Deadlines are tight, expectations are high, and incomplete information can quickly create downstream challenges.
Recent developments from the Federal Trade Commission make that early understanding even more important. New FTC instructions governing the use of technology-assisted review (TAR) and AI-assisted review in certain investigational subpoenas, including merger Second Requests, impose significantly more detailed requirements around collection, review methodology, validation, and disclosure. Among other things, the instructions contemplate collection from all custodians being complete before TAR or AI model training begins.
Before collections begin, teams need a clear understanding of their data environment. Knowing where potentially responsive information resides, how it is stored, who controls it, and what limitations may exist can significantly improve efficiency, reduce risk, and support defensible decision-making throughout the response.
Many organizations assume they have a solid understanding of where their business information resides.
In reality, the modern data landscape is often far more complex.
Responsive information may be scattered across:
Email platforms
Without an accurate picture of these sources, teams risk overlooking relevant information, collecting unnecessary data, or creating costly delays later in the process.
That risk becomes particularly important when TAR or AI-assisted review is being considered. Under the new FTC instructions, parties may need to provide information about the types of documents included in the review population, file types excluded for special handling, ephemeral and text messages, and linked or modern attachments. The instructions also state that model training should not begin until collection from all custodians is complete and the FTC has had an opportunity to review the required disclosures.
The practical takeaway is straightforward: collection strategy and review strategy can no longer be treated as completely separate phases.
The most effective Second Request responses begin with understanding the data before designing either one.
One of the most valuable early exercises is meeting with the individuals who understand the organization's technology environment best.
An IT interview helps legal and response teams identify:
These conversations often uncover critical details that may not be obvious at the outset of a matter.
For example, a business unit may primarily use email for communication, but key decisions could also reside in collaboration platforms, cloud storage applications, or archived systems that are not immediately apparent to counsel.
Modern attachments, ephemeral messaging, text messages, and other nontraditional data sources deserve particular attention. The FTC's updated instructions specifically address how certain file types are handled when TAR or AI-assisted review is contemplated, making it increasingly important to understand these sources before the review population is finalized.
Understanding these systems early allows teams to make informed decisions about scope, collections, review strategy, and resource allocation.
While IT teams can explain systems and infrastructure, custodians provide a different perspective.
They explain how data is actually created, stored, and shared in day-to-day operations.
Structured custodian interviews can help teams understand:
These interviews frequently uncover previously unidentified data sources, repositories, or custodians that would not have been identified through technical analysis alone.
In many cases, they also reveal practical workflow details that help focus collections and reduce unnecessary review volumes.
For matters in which TAR or AI-assisted review may be used, identifying the complete custodian population early takes on additional significance. If model training cannot begin until collection from all custodians is complete, discovering a material custodian late in the process could affect not only collection but also the review timeline.
Once key stakeholders have been interviewed, organizations can begin building a comprehensive view of their information landscape.
Data mapping provides a structured way to document:
Rather than treating collections as a series of disconnected activities, data mapping helps create a clear relationship between custodians, systems, and responsive information.
This visibility becomes increasingly important as matters evolve and new custodians or repositories are identified.
A well-maintained data map also supports defensibility by demonstrating the rationale behind collection decisions and helping teams document their response efforts.
Once teams understand their data environment, they can move into collections with greater confidence.
At this stage, defensibility becomes critical.
Organizations should be prepared to document:
As new custodians, repositories, or responsive data sources emerge, collection plans may need to evolve.
Maintaining detailed documentation helps ensure those changes are captured and can be explained if questions arise later.
That documentation can also inform downstream conversations with regulators about review methodology. The FTC's newer instructions call for extensive disclosures when TAR or AI-assisted review is used, including the technology and workflow, model information, training and validation methodologies, review protocols or prompts, and certain statistical measures.
A defensible collection process establishes a reliable foundation for processing, review, quality control, and production activities that follow.
Consider a scenario where a key executive is initially identified as a primary custodian.
Early planning suggests collections will be limited to email and network documents.
However, during interviews, the response team discovers the executive also relies heavily on collaboration platforms, maintains files in multiple cloud repositories, uses a mobile device for business communications, and has historical information stored in an archive environment. The interview process also identifies an additional custodian with significant involvement in the relevant business activities.
Without those conversations, critical information could have been overlooked or discovered much later in the process, creating delays, supplemental collections, and additional review costs.
If the team was also planning to employ TAR or AI-assisted review, the newly identified custodian could have broader consequences. Under the new FTC instructions, completing custodian collections can be a prerequisite to beginning model training, meaning late discoveries may affect the review schedule itself.
By identifying these facts early, the team can adjust its strategy before collection and review workflows are fully underway.
Review workflows, analytics, quality control, and production strategies all depend on one thing: understanding the data first.
Organizations often focus on collection as the starting point of a Second Request response. In reality, the most effective responses begin earlier.
As regulatory expectations around TAR and AI-assisted review become more prescriptive, the relationship between data identification, collection, review methodology, validation, and production is becoming even tighter.
By investing time in IT interviews, custodian interviews, data mapping, and collection planning, legal teams can make better decisions about scope, prioritize the right information, and build a stronger foundation for the work ahead.
Before collecting terabytes of data, consider these questions:
The answers can have a significant impact on the efficiency, defensibility, and overall success of a Second Request response.
Successful Second Request responses depend on more than collecting data. They depend on understanding where responsive information resides, how it can be collected, and how collection decisions will affect downstream review and production workflows.
Recent FTC developments make that connection even clearer. For teams considering TAR or AI-assisted review, early visibility into custodians, systems, file types, and collection completeness may directly affect when review can begin and how the methodology must be documented and validated.
Array helps legal teams navigate complex data environments, develop defensible collection strategies, and build efficient workflows for large-scale regulatory matters.
Contact Array to discuss your Second Request (SIR) response strategy.