Why volume matters
- 73%
- of production cost in the typical cases studied went to document review
- RAND Corporation, 2012, study of 57 cases
- 57
- cases informed the production-cost study
- RAND Corporation, Where the Money Goes
- 4.9/5
- average client rating across delivered projects
- Clutch, verified reviews
RAND found that review for relevance, responsiveness, and privilege accounted for about 73% of eDiscovery production costs in the cases it studied. The report also warns that its sample does not represent every litigant or matter. The useful conclusion is narrower: when review dominates the bill, reducing the population before linear attorney review deserves a measured test. Read the RAND report.
Custom review automation earns its place on large, repeatable matters.
The review protocol and evidence requirements matter more than the novelty of the model.
A fit01Your matters contain enough documents that linear review drives a material share of cost or deadline risk.
02Counsel can define relevance, privilege, issue codes, and a validation method before processing begins.
03Your existing review platform cannot deliver the classification, quality control, or reporting the matter requires.
Not a fit01The population is small enough for a qualified reviewer to finish faster than a pilot could be configured.
02The matter team has not agreed what relevance or privilege means for this production.
03A standard feature in your current eDiscovery platform already solves the problem cleanly.
Review layer
What the system controls
01Relevance and issue classification
Counsel's written criteria become review fields and confidence thresholds. Strong candidates move to the right queue; uncertain documents stay visible for attorney coding. Issue tags organise the population around the questions that matter to the case.
02Duplicates, email threads, and families
Exact duplicates, near-duplicates, email threads, and attachments are grouped so reviewers see context without repeating the same decision across copies. Family rules stay configurable because attachments and parent emails can carry different review obligations.
03Privilege flags and review workflow
The system flags possible attorney-client, work-product, and other defined privilege signals. It does not make the final privilege call. Reviewers confirm the result, record the reason, and produce the fields needed for a privilege log.
04Validation and audit history
Held-out samples measure recall and elusion against attorney-coded decisions. The system records reviewer decisions, coding changes, thresholds, model versions, and exports so the team can reconstruct how a document reached its final status.
Linear review vs AI-assisted review
| Linear review | AI-assisted review |
|---|
| Starting point | Reviewers open documents in sequence | Duplicates are grouped and documents are ranked before review |
| Attorney time | Spread across the full population | Directed to uncertain, responsive, and high-value material |
| Quality control | Often sampled after coding | Sampling and disagreement checks run throughout review |
| Completion | Ends when the queue is empty | Ends when counsel accepts the documented validation result |
| Best fit | Small or unusual populations | Large populations with stable review criteria |
How it works
From review protocol to one live matter
- Phase 1
01Define the protocol
Agree the matter scope, relevance criteria, privilege categories, issue list, validation method, and completion rule with counsel. The system cannot rescue an ambiguous protocol later.
- Phase 2
02Test a labelled sample
Run representative documents through the proposed approach and compare the results with attorney coding. This exposes OCR, language, and classification problems before they reach the full population.
- Phase 3
03Process and review
Group duplicates, score the remaining population, and route uncertain or high-value documents to the right reviewers. Quality checks watch for inconsistent coding while the matter is still active.
- Phase 4
04Validate and close
Measure recall and elusion on held-out samples, document exceptions, and preserve the full decision and model history. Counsel owns the decision that the review is complete.
- A vague issue list
- If reviewers interpret the same criterion differently, the training signal becomes noise. Counsel needs to settle examples and edge cases before wider coding.
- A convenient sample
- Testing only clean, obvious documents hides the failures that matter. The sample needs scans, long threads, attachments, rare issues, and likely privilege.
- A black-box completion call
- A confidence score is not a defensibility record. The matter needs an agreed validation method, measured results, and a named decision-maker.
Pilot scope
Start with one matter at $30,000.
We use a representative document sample to prove the protocol, integration path, and validation method before the full population is processed.
If your current review platform already covers the workflow, we will say so. Custom work is justified only when the missing review layer costs more than it solves.
Starting investment
Starts at $30,000
A focused first release usually takes 10 to 14 weeks. Document volume, OCR quality, integrations, and validation requirements move the number.
Scoped before development
The first matter, sample, integrations, and acceptance checks are written down before development begins. Changes are priced and agreed before they enter the work.
Attorney-controlled completion
The system can measure and document the review. Counsel retains the final privilege, production, and completion decisions.
