Gavel provides a cloud‑based e‑discovery platform that lets lawyers upload, process, and search large document sets with AI‑assisted tools while keeping all decisions under human control.
Funding
Funding not disclosed
Founders
Product
Problem
Law firms and barristers must process massive volumes of documents for e‑discovery, which involves time‑consuming tasks such as virus scanning, format conversion, OCR, metadata extraction, deduplication, and manual relevance coding. Existing platforms often require extensive installation, per‑seat licensing, and place AI decisions outside lawyer oversight, creating compliance and security concerns.
Solution
Gavel offers a cloud‑based e‑discovery platform that lets lawyers upload any document format directly through a browser, where the system automatically scans, converts, OCRs, extracts metadata, deduplicates, and indexes files. Lawyers use a plain‑English AI query builder to generate deterministic searches, which they must review and approve before execution. AI also proposes categorizations, dates, and descriptions for documents, marking each suggestion as editable so that lawyers retain final responsibility for relevance and privilege decisions. Predictive review leverages manual coding to prioritize uncertain or likely‑relevant documents, accelerating review without removing lawyer control. All data and AI processing are hosted in Australia, with full audit trails and security certifications.
Target Audience
Gavel is designed for Australian law firms, barristers, and in‑house legal teams that require secure, compliant e‑discovery tools for large‑scale document review and litigation support.
Features
- Browser‑based upload supporting large files, multiple formats, and tranche organization with pre‑submission review
- Automated background pipeline: virus scanning, native conversion, OCR, text and metadata extraction, deduplication, and searchable indexing
- AI query builder that translates plain‑English requests into structured queries, requiring lawyer approval before running
- AI‑assisted coding for categories, dates, and descriptions, with each output marked, editable, and under lawyer oversight
- Predictive review model that learns from manual coding to surface uncertain or high‑relevance documents first
- Full review workspace with true (burned‑in) redaction, inline drafting tools, and export of discovery sets with hyperlinked indexes
- Unlimited user access per matter with no per‑seat fees, audit‑ready append‑only logs, and secure client upload portal
- Australian‑hosted infrastructure, ISO 27001 and SOC‑2 compliance, and data never used to train external AI models