Discovery operations

Discovery Document Triage: Reduce Review Load Without Losing Traceability

A discovery preparation workflow for inventories, processing exceptions, review batches, and source-linked coding that leaves legal decisions with counsel.

By WhichAI. Published 2026-07-12. Updated 2026-07-12.

Methodology: Editorial synthesis of workflow design patterns and implementation constraints. Public control references provide context, not proof of a deployment or legal advice. Where a versioned evidence pack appears, its evidence class, method, and limitations govern what the artifact can support. Read the full method. Report a correction.

Built for

Litigation teams, discovery counsel, legal operations leaders, and review managers

The decision

Decide where machine-assisted grouping and prioritization can reduce review preparation without weakening preservation, privilege, or production controls.

Answer first

Use automation to inventory sources, detect processing failures, organize review batches, and surface likely patterns. Keep responsiveness, privilege, redaction, withholding, and production decisions with authorized legal reviewers.

WhichAI Solutions diagnostic

Bring this operating problem to the diagnostic

Use Solutions when discovery preparation spans custodians, collection tools, processing platforms, review teams, privilege controls, and production approvals with material legal risk.

Open the diagnostic

The capacity leak

What the team is doing before anyone calls it a systems problem

Headcount pressure rarely starts with one giant task. It starts when ordinary work is split across inboxes, tabs, handoffs, and undocumented judgment calls. These are the signals to map first.

SIGNAL 01

Collections arrive from several custodians and systems with inconsistent metadata and processing histories.

SIGNAL 02

Review teams spend substantial preparation time deduplicating, threading, grouping, and locating obvious exceptions.

SIGNAL 03

A priority score can influence review order without showing which source signals produced it.

SIGNAL 04

Coding changes and production decisions are difficult to reconstruct across tools and review waves.

The implementation

The system should prepare the decision, not pretend the decision disappeared

A complete implementation connects the intake, context, transformation, review, and record. The output of one stage becomes the controlled input to the next. A human owns the exceptions and the final consequence.

Stage	Current drag	System responsibility	Human responsibility	Evidence kept
1. Collection inventory	Sources and custodians are tracked outside the review workspace.	Create an inventory of collections, custodians, date ranges, transfer records, and preservation status.	Counsel defines scope and resolves preservation or collection questions.	Source identifier, custodian, collection method, checksum, and transfer log.
2. Processing control	Failed extraction, corrupt files, archives, and unsupported formats are handled ad hoc.	Process copies, detect failures, and route technical exceptions without altering originals.	Approve exception handling and investigate gaps that may affect completeness.	Original checksum, processing version, failure reason, and resolution.
3. Review preparation	Documents are assigned in large undifferentiated batches.	Group threads, near-duplicates, and agreed categories while retaining document-level source links.	Approve batching logic and ensure prioritization does not exclude required review.	Group membership, ranking inputs, batch rule, and assigned reviewer.
4. Legal coding	Review decisions are entered without consistent rationale or quality sampling.	Present document context, family relationships, and prior coding in a controlled review interface.	Make responsiveness, privilege, redaction, and issue decisions.	Reviewer, code, rationale, source document, and review timestamp.
5. Production record	The path from collection to produced item is fragmented across exports.	Assemble authorized production candidates and preserve lineage through any approved transformation.	Counsel approves production, withholding, redaction, and final release.	Production number, source ID, transformation log, approval, and release record.

What the human keeps

The goal is not zero humans. It is zero avoidable preparation around the judgment only a responsible owner should make.

Counsel defines collection and review scope and makes responsiveness, privilege, redaction, and production decisions.
Discovery specialists resolve processing failures, source gaps, family relationships, and technical exceptions.
Review managers run quality sampling and investigate systematic coding or prioritization errors.

Controls before volume

A workflow is not ready because the happy path worked once. It is ready when access, review, fallback, and evidence are explicit.

Preserve immutable originals and chain-of-custody records before processing or enrichment.
Do not use a priority score to remove documents from required legal review.
Retain document-level reasons and source signals for grouping and prioritization.
Require authorized counsel approval before any production export is released.

Control references

National Institute of Standards and Technology: Privacy Framework Cybersecurity and Infrastructure Security Agency: Secure by Design

The scorecard

Measure capacity, not activity

A system can produce more messages and still make the operation worse. Measure movement through the workflow, the quality of review, and the load that still reaches a person.

Preparation hours per batch

Human time spent processing exceptions, organizing context, and making a batch review-ready.

Processing exception closure

Share of corrupt, unsupported, or incomplete items with a documented resolution.

Reviewer correction rate

Share of sampled legal coding decisions changed through quality review.

Production lineage

Share of produced items traceable to original collection, transformations, coding, and approval.

What a fake implementation looks like here

These patterns create an AI demo while leaving the labor, risk, and accountability in the same place.

Treating semantic similarity or model relevance as a legal responsiveness decision.
Dropping failed or unsupported files from the population without an exception record.
Breaking email families or document context to improve batch efficiency.
Producing transformed documents that cannot be traced to immutable collected originals.

Two ways to act

Use the path that matches the decision

WhichAI Solutions

The workflow is becoming a company problem.

Use Solutions when discovery preparation spans custodians, collection tools, processing platforms, review teams, privilege controls, and production approvals with material legal risk.

Bring one bottleneck. We map the work under it, separate consequential judgment from mechanical drag, and decide whether the next move is a hire, a tool, or a rebuild.

See company solutions

Task-specific workflow brief

Plan this recurring task.

Start with this task draft, then complete the three-question brief:

Design a discovery document triage workflow with collection inventory, immutable originals, processing exceptions, review batches, document-level source links, counsel coding, quality sampling, and production lineage.

Choose a paid plan after reviewing your brief. WhichAI creates a plan and does not set up tools or accounts.

Start the brief

Questions

What operators ask before they build

Can AI decide which discovery documents are responsive?

This workflow does not assign final responsiveness, privilege, redaction, or production decisions to AI. It prepares review and prioritizes work under counsel-defined controls.

What traceability should a review batch preserve?

Keep the original document ID, collection source, family relationships, processing history, grouping reason, ranking inputs, reviewer coding, and quality-review outcome.

How can a team test review-load reduction safely?

Run a bounded shadow review, preserve the full review population, compare preparation hours and reviewer corrections, and examine errors by document type and issue.

Primary references

Controls should come from the specific operating environment

These are broad public control references, not article-specific evidence, vendor endorsements, or legal advice. Validate the current rules, contracts, system configuration, and organization-specific risk before deployment.

National Institute of Standards and Technology

Privacy Framework

A framework for identifying and managing privacy risk in products and operations.

Accessed 2026-07-14

Cybersecurity and Infrastructure Security Agency

Secure by Design

Security principles for making systems safer by default and reducing avoidable customer burden.