Unnamed Fortune 500 Company

As a Fortune 500 construction company managing multi-billion dollar infrastructure and commercial projects, the in-house legal team faced exponentially growing data volumes tied to litigation — including mobile phone records, video files, and CAD and design files. Delay and disruption claims routinely triggered demands for every record across a project, producing document sets in the hundreds of thousands. The company had fully outsourced e-discovery to third-party vendors and outside counsel, leaving the legal team without visibility into the process, without control over work product, and unable to surface decision-critical information efficiently. The result was escalating external spend and slower dispute resolution, undermining the company's stated priority of early case settlement.

The company brought e-discovery entirely in-house using Everlaw's litigation platform, centralizing all matters onto a single system and replacing fragmented legacy databases and vendor relationships. The core of the transformation was Everlaw's machine learning-powered document prioritization, including Multi-Matter Models — reusable trained AI that applies predictive coding across successive cases, learning patterns with each matter to reduce human review time over time. The team layered in Data Visualizer and Communications Visualizer tools to map thematic clusters and surfaced key party communications as interactive graphs, enabling earlier identification of critical documents. Work was restructured into phases: the in-house team handles early culling and first-pass review, then hands off a targeted, pre-analyzed subset to outside counsel for issue-specific work.

In one major matter, the approach cut estimated e-discovery costs by 70% — from $260,000 down to $75,000 — while delivering faster turnaround and higher-quality work product:

• Starting corpus of 700,000 records culled to 400,000 after initial relevance filtering
• 400,000 documents completed in approximately three weeks by a small in-house team
• Outside counsel received 200,000 records, half the volume that would have gone out under the legacy model
• 2,500 critical documents pre-identified for immediate review, accelerating case strategy from day one

The AGC noted the savings were roughly one-fifth of the previous per-volume cost estimate. Year-over-year, the team has achieved more favorable resolutions in less time.

• Insourcing e-discovery fundamentally shifts control: owning the platform means decision-critical information is surfaced by the legal team, not filtered through vendor priorities.
• AI models compound in value across matters: Multi-Matter Models improve with each case, meaning the efficiency gains from predictive coding increase over time rather than resetting.
• Phase the work, don't batch it: separating in-house culling from outside counsel review maximizes the value of both and prevents unnecessary cost escalation at the handoff stage.
• Early investment in data visualization pays off during review: mapping document relationships visually reduces the time needed to identify relevant clusters in large, heterogeneous datasets.
• Templatized workflows are a construction-specific advantage: the industry's repeatable litigation patterns (delay, disruption, defect claims) make first-pass review automation especially tractable.