A small law firm engaged Avalon to support a large-scale eDiscovery matter involving a rapidly expanding document population and aggressive production deadlines. Rather than relying on a traditional linear review, the firm wanted a faster, more defensible approach that could identify relevant documents, manage privilege, and address sensitive financial information and personally identifiable information (PII) while maintaining consistency throughout the review process.
Avalon partnered with counsel early in the matter to help develop an ESI protocol built around Relativity aiR for Review. Beyond implementing the technology, Avalon designed and validated a complete AI-driven workflow that aligned with the firm’s case strategy, reduced manual review, and created a practical process for review, redactions, privilege analysis, quality control, and production.
The Challenge
What began as an anticipated review of approximately 25,000 documents quickly expanded into a review universe of nearly 400,000 documents. A traditional first-pass review would have required significant time and cost while making it difficult to meet production deadlines.
The matter also required a defensible review methodology that could be validated and communicated to opposing counsel. In addition to determining relevance, the team needed to consistently identify privileged material, financial records, and PII requiring redaction, all while maintaining accuracy and control across a rapidly growing dataset.
The Strategy
Avalon replaced the traditional first-pass review with a customized Relativity aiR for Review workflow. Working closely with counsel, the team developed and refined AI prompts that reflected the firm’s relevance criteria, then validated the results against attorney coding decisions before applying the workflow across the entire document population.
Once validated, aiR categorized documents into Relevant, Borderline, and Non-Relevant groups, allowing reviewers to focus only on documents requiring legal judgment instead of manually reviewing every document.
Avalon then built downstream workflows around those results, incorporating privilege review, PII detection, financial document identification, redactions, and quality control. By aligning every stage of review with the aiR outputs, the team was able to keep review and production moving simultaneously while maintaining a consistent, defensible process.
The Results
The aiR for Review workflow reduced the manual review population from approximately 400,000 documents to about 40,000 requiring attorney review, allowing reviewers to concentrate on higher-value work such as privilege determinations, redactions, and quality control.
The team successfully supported multiple production rounds, ultimately producing approximately 200,000 documents (roughly 500,000 pages). Because review activities were prioritized using aiR, production progressed in parallel rather than waiting for a complete linear review.
The workflow also delivered strong defensibility metrics, achieving 99.4% recall, 0.4% elusion, and 77.0% precision. The engagement demonstrates how combining Relativity aiR for Review with Avalon's workflow design and litigation support expertise can reduce review costs, accelerate timelines, and produce consistent, defensible results at scale.
