RFI-ICE_EDS_OCT2026.pdf

RFI-ICE_EDS_OCT2026

REQUEST FOR INFORMATION (RFI)

Enterprise Software-as-a-Service (SaaS) Electronic Discovery (eDiscovery) Platform Including Super Intelligence (SI) and Agentic SI capabilities

Issued By:

U.S. Immigration and Customs Enforcement (ICE) Office of the Principal Legal Advisor Knowledge Management Division

Issue Date:

October 19, 2026

Response Due Date:

October 28, 2026 at 1:00 p.m. Eastern Time (ET)

1.0 Purpose and Disclaimer

This is a RFI is for the ICE, Office of the Principal Legal Advisor (OPLA), regarding an Enterprise SaaS eDiscovery Platform including SI and Agentic SI capabilities.

THIS IS A REQUEST FOR INFORMATION (RFI) ONLY. This is not a solicitation announcement. This RFI is issued solely for market research purposes in accordance with Revolutionary Federal Acquisition Regulation (FAR) Overhaul (RFO) Part 10. This RFI does not constitute a Request for Proposal (RFP), Request for Quotation (RFQ), Invitation for Bid (IFB), or other solicitation. The Government is under no obligation to issue a solicitation, award a contract, or reimburse respondents for any costs incurred. All costs associated with responding to this RFI will be solely at the interested party's expense. Responses will be used to assist the Government in acquisition planning, market research, and requirements development. The information provided in this RFI is subject to change and is not binding to the Government.

Proprietary Information: All submissions become the property of the Government and will not be returned. If submitting proprietary information, clearly mark “PROPRIETARY” on every applicable sheet and segregate such information from other portions of the response (e.g., via attachment). Respondents are responsible for adequately marking proprietary information in their response. All information received in response to this RFI that is marked “Proprietary” will be handled accordingly.

2.0 Agency Overview and the Office of Principal Legal Advisor The Office of the Principal Legal Advisor (OPLA), a component within the ICE federal agency, is the largest legal program in the Department of Homeland Security (DHS), with over 2300 attorneys and 350 support personnel. By statute, OPLA serves as the exclusive representative of DHS in immigration removal proceedings before the Executive Office for Immigration Review, litigating all removal cases including those against criminal aliens, terrorists, and human rights abusers.

OPLA also provides a full range of legal services to ICE programs and offices. OPLA provides legal advice and prudential counsel to ICE personnel on their customs, criminal, and immigration law enforcement authorities, the Freedom of Information Act and Privacy Act, ethics, legal liability under the Federal Tort Claims Act, and a range of administrative law issues, such as contract, fiscal, and employment law. OPLA represents the Government before the Merit Systems Protection Board, the Equal Employment Opportunity Commission, and the Board of Contract Appeals. OPLA attorneys provide essential support to the Department of Justice in the prosecution of ICE cases and in the defense of ICE’s authorities in federal court.

In addition to its Washington, D.C. location, OPLA has offices in more than 60 locations throughout the United States. ICE OPLA handles a large volume of cases and associated data. OPLA handles all aspects of the electronic discovery lifecycle using a SaaS based electronic discovery platform.

3.0 Purpose of Request for Information

ICE is conducting market research in accordance with RFO Part 10 to identify commercially available, cloud-native SaaS eDiscovery platforms capable of supporting the enterprise-wide litigation, investigations, administrative proceedings, Freedom of Information Act (FOIA) litigation, regulatory matters, congressional inquiries, and other legally, regulatorily required discover activities or compliance-related activities.

ICE anticipates a continuing need for a secure, scalable, cloud-based eDiscovery platform that supports the full Electronic Discovery Reference Model (EDRM), from information governance and legal hold through identification, preservation, collection, processing, review, analysis, and production.

As the volume, variety, and complexity of electronically stored information (ESI) continue to increase, the Government seeks to better understand the capabilities of commercially available platforms that improve the efficiency, defensibility, security, and quality of eDiscovery operations. In particular, the Government is interested in solutions that reduce manual effort, streamline workflows, enhance collaboration, improve data analysis, and support timely decision-making while maintaining appropriate human oversight and compliance with applicable legal and regulatory requirements.

The Government recognizes that recent advances in SI, including generative SI, machine learning, predictive analytics, and emerging agentic SI capabilities, have the potential to augment traditional eDiscovery workflows. Accordingly, the Government seeks information regarding how these technologies have been incorporated into commercially available eDiscovery platforms to enhance, automate, or assist users throughout the EDRM lifecycle. SI capabilities should complement and enhance core platform functionality rather than replace established eDiscovery processes.

The Government is interested in understanding both currently available capabilities and the maturity of emerging technologies, including autonomous workflow orchestration, intelligent task automation, contextual analysis, natural language interaction, explainable SI, and configurable human-in-the-loop controls.

Information received in response to this RFI will assist the Government in:

• Assessing the current commercial marketplace for enterprise SaaS eDiscovery platforms.

• Identifying mature commercial capabilities across all phases of the EDRM.

• Understanding the availability, maturity, and practical application of SI-enabled capabilities, including agentic SI, within commercial eDiscovery platforms.

• Evaluating approaches to improving efficiency, defensibility, scalability, and user productivity through automation and intelligent assistance.

• Identifying capabilities that support secure operation within a federal cloud environment, including compliance with applicable federal cybersecurity requirements.

• Informing acquisition planning, requirements development, independent government cost estimating, and future solicitation strategies.

4.0 Technical Requirements and Objectives

4.1 Platform Overview and Desired Solution

The Government seeks a fully integrated, cloud-native SaaS eDiscovery platform capable of supporting the complete EDRM lifecycle without requiring multiple independent products. The solution must support matters of varying sizes and complexity, from routine investigations to large-scale, enterprise-wide litigation. The platform shall minimize manual processes, improve user productivity, reduce review costs, and support defensible discovery practices.

4.2 Technical & Performance Objectives

The solution must deliver reliable performance across all EDRM phases. Key objectives include:

Performance: Reduce document review time, support large data volumes, maintain responsiveness under heavy workloads, and increase reviewer consistency.

Operational Efficiency: Increase automation, reduce manual effort, improve search effectiveness, and accelerate production timelines.

Security & Compliance: Maintain an active FedRAMP Authorization (Moderate or High). Ensure transparency, auditability, and appropriate human oversight for all SI-assisted actions.

4.3 Functional Scope (EDRM)

The platform must provide comprehensive functionality across the following areas:

Information Governance & Preservation: Legal hold, tracking, notifications, and risk monitoring.

Collection & Processing: Multi-source ingestion (cloud, email, collaboration platforms, endpoints), OCR, metadata extraction, deduplication, email threading, and language identification.

ECA & Analysis: Data clustering, search refinement, case analytics, entity extraction, communication/relationship analysis, and chronology/timeline generation.

Review & Production: TAR/Continuous Active Learning, SI-assisted coding/summarization, PII/Privilege detection, redaction, production validation, and reporting.

4.4 Advanced SI and Agentic SI Objectives

The Government seeks to leverage SI beyond traditional predictive coding to include intelligent workflow orchestration. Key objectives include:

Intelligent Automation: Autonomous execution of multi-step discovery tasks, SI-generated summaries, and automated quality assurance.

Contextual Analysis: Semantic searching, natural language querying, relationship mapping, and automatic issue identification.

Agentic Capabilities: The Government is interested in SI agents that can coordinate complex workflows, recommend investigative actions, and adapt based on user feedback, all while operating within administrator-defined policy boundaries and configurable human-in-the-loop (HITL) approval checkpoints.

5.0 White Paper Content

White papers shall be submitted in accordance with the following requirements. While there is no page limit, submissions should be concise and focused. At a minimum, interested parties shall provide the following:

• Administrative Information: A cover letter (1-2 pages) including:

o Company Name, UEI, CAGE Code, and HQ location.

o Business Size (under NAICS 541512).

o Point of Contact and Website.

o Available IT contract vehicles.

o Indication of interest in a One-on-One session.

o Executive Summary.

• Company Profile: Years in business, relevant federal experience, FedRAMP authorization level (including sponsoring agency and authorization date), cloud deployment model, and hosting locations.

• Capability Narrative: Description of current capabilities and planned enhancements. Please distinguish between General Availability (GA) vs. roadmap items, and configuration vs.

customization details. Include information regarding third-party product usage and licensing/pricing models.

• Artificial Intelligence & EDRM Responses: Provide structured responses to the functional requirements outlined in sections 8.1 and 8.2. Clearly distinguish between GA, limited availability, preview/beta, and planned capabilities. Respondents must clearly indicate if features require optional licensing or third-party integration and provide sufficient detail to evaluate maturity.

6.0 Industry One-on-One Sessions

One-on-one sessions are not currently scheduled; however, the Government may invite select vendors to participate in 1-2 hour sessions based on the review of submitted white papers. Participation is voluntary, non-marketing focused, and will have no bearing on future contract awards. Only vendors who submit a responsive white paper will be considered. The Government reserves the right, at its sole discretion, to select vendors for these sessions to demonstrate their products or services.

7.0 Submission Instructions

• Format: White papers shall be submitted in Adobe PDF format, using 12-point Times New Roman font, single spacing, and 1-inch margins.

• Deadline: Please submit responses no later than 1:00 p.m. ET, on October 28, 2026. Early submissions are encouraged and welcomed.

• Q&A Policy: Please note that there will be no formal Question and Answer (Q&A) period for this RFI.

All information required to respond is contained within this document.

• Submission Method: Responses should be submitted via email to OAQ-ITD-SD-requirements@ice.dhs.gov.

• Subject Line: Please use the following subject line: ICE eDiscovery SaaS – RFI Response – [Company Name].

8.0 Technical Response Requirements

Vendors shall structure their white paper responses to address the following requirements.

Respondents shall provide narrative descriptions and, where applicable, quantitative metrics, benchmarks, or customer examples for all functional areas. When addressing SI and EDRM capabilities, responses must explicitly distinguish between:

• Current Availability: Generally Available (GA), Limited Availability, Preview/Beta, or Planned (with timelines).

• Licensing/Dependencies: Whether capabilities are standard, require optional licensing, or require third-party product integrations.

Respondents should describe how their solution is designed to: Reduce document review time, improve reviewer consistency, increase automation, reduce manual effort, improve search effectiveness, accelerate production timelines, support large data volumes, maintain responsiveness under heavy workloads, improve administrative efficiency, and enhance overall user productivity. Where available respondents are encouraged to provide quantitative metrics benchmarks or customer examples, demonstrating measurable improvements. Responses should include sufficient detail and may reference product literature, white papers, or technical documentation to enable the Government to evaluate the maturity and applicability of the proposed capabilities.

8.1 SI, Agentic SI, and Governance Response Structure

For every SI and Agentic SI capability described, respondents shall address the following:

mailto:OAQ-ITD-SD-requirements@ice.dhs.gov mailto:OAQ-ITD-SD-requirements@ice.dhs.gov

• Classification & Availability: Indicate if the feature is GA, limited availability, beta/preview, or planned. State whether it is part of the standard commercial offering or requires optional licensing/separate modules.

• Architecture & Models: Describe the underlying architecture, identify SI models used (proprietary, open-source, or hybrid), latency and resource utilization, and explain the selection methodology for specific tasks.

• Operational Integration: Detail how SI/Agentic capabilities are integrated into the EDRM workflow. Address natural language interfaces, semantic search, generative SI, automated summarization, and SI-assisted coding/classification.

• Agentic Capabilities: For agentic features, describe the ability to execute multi-step workflows, explain actions taken, coordinate agents, monitor for events, prioritize activities, and operate within defined policy boundaries, generate task plans from natural language, and handle confidence thresholds (e.g., suspending execution when thresholds are not met).

• Security & Privacy: Describe tenant and prompt isolation, data segregation, encryption (prompts, responses, and SI content), protection against prompt injection/unauthorized access, data residency/sovereignty, and role-based access controls. Respondents shall clearly state whether customer data, prompts, or documents are used to train foundation models.

• Explainability & Auditability: Describe how SI recommendations are explained, the availability of confidence scores/supporting rationale, and how SI activity is captured in audit logs and how the platform generates comprehensive, explorable audit logs (including user attribution and prompt/response logging) for all SI-assisted activities. Detail the ability to reconstruct SI-assisted workflows for legal or regulatory review.

• Responsible SI & Oversight: Describe human-in-the-loop (HITL) mechanisms, user override capabilities, model red-teaming/vulnerability testing, confidence scoring methodologies, methods for reducing hallucinations, and procedures for identifying/mitigating model bias.

• Governance Framework: Describe organizational SI governance policies, risk management practices, interoperability/portability, model lifecycle management (validation, version control, testing), and customer notification procedures for significant model changes.

8.2 EDRM Functional Response Structure

For each phase of the EDRM—Identification, Preservation, Collection, Processing, ECA, Review, Analysis, and Production—respondents shall:

• Describe Functionality: Detail how the platform supports specific activities (e.g., legal hold management, chain of custody, OCR, deduplication, redaction, production validation) using native capabilities and optional modules.

• Highlight Enhancements: Explicitly describe how SI and Agentic SI features are integrated into the core workflow and leveraged to improve efficiency, accuracy, and defensibility within that specific EDRM phase.

• Identify Dependencies: Explicitly note if any functional area requires third-party products or integrations to achieve full capability.

• Optional Modules: Any functionality dependent on optional licensing.