Work / 2023
Cloud-native data collection at Everlaw
Modernising how legal teams get data into an ediscovery platform — from file uploads to direct collection from wherever the data actually lives.
Context
Everlaw is a cloud-based ediscovery platform — legal teams use it to collect, process, and review data for litigation and investigations. When the platform was built, data collection meant uploading files from local drives or on-premise servers. By the time I was working on it, the data teams needed for legal review increasingly lived in cloud SaaS tools: Zendesk, Slack, cloud storage, ticketing systems.
The platform hadn’t kept up. Getting data from these sources required service providers to extract and reformat it manually before it could enter Everlaw. That was slow, error-prone, and an unnecessary step in a workflow that was supposed to be fast.
Problem
Two overlapping problems:
Collection: No way to pull data directly from cloud SaaS sources. Every new source required a manual extraction step before data could enter the platform — adding time, cost, and a point of failure in the chain of custody.
Visibility: Once data was in flight, tracking progress was fragmented. Users had to check multiple views to understand what was uploading, what had failed, and where things stood for large collections. 77% of incoming platform data passed through tooling that was difficult to monitor.
What I shaped
Two parallel efforts, designed to complement each other:
Zendesk API integration — direct collection from Zendesk without manual extraction. Legal teams could search, collect, and review Zendesk ticket data in the same workflow as everything else in the platform. No service provider manipulation required.
Ingestion monitoring redesign — consolidated progress reporting into a single interface covering local and cloud data collections. The goal was to make the state of any collection obvious at a glance, and to surface failures early enough to act on them.
Key decisions and trade-offs
Zendesk as the first cloud-native integration. Several cloud sources were on the list. Zendesk was the right starting point: high demand from existing customers, a well-documented API, and a clear substitution for a painful manual workflow. Starting here also let us build the integration pattern before extending it.
Treat monitoring as a product, not an afterthought. The ingestion redesign wasn’t glamorous, but fragmented monitoring was real friction for users running large collections. Better visibility reduces support load and builds trust in the platform’s reliability.
Design the integration layer to extend. Zendesk wasn’t the only cloud source on the roadmap. The integration architecture was built so that adding new sources followed a consistent pattern, not a one-off build each time.
What launched
Zendesk integration shipped — legal teams could directly search, collect, and review ticket data within their existing Everlaw workflow, eliminating the need for service providers to extract and manipulate data into review formats first.
Ingestion monitoring redesign improved visibility across 77% of incoming platform data, consolidating reporting that was previously spread across multiple views.
What I’d do next
Extend the cloud-native integration pattern to other high-frequency sources: Slack, Google Workspace, cloud storage. The Zendesk work established the pattern; the value compounds as more sources get added.