Discernis Raises $2.5M for AI-Native Legal Discovery

Discernis is financing a bet that legal AI should understand the whole record before a lawyer has to defend an answer. The New York City company announced a $2.5M Seed round on August 25, 2026, led by Newfund Capital with Triple Impact Capital, Remarkable Ventures, and C2 Ventures participating.

The capital will support continued product development and expansion across large law firms and corporate legal departments. The larger question is whether whole-collection analysis can move legal AI beyond persuasive retrieval and into work where completeness, privilege, security, and explainability have to survive an adversarial process.

Discernis says its software reads every document in a collection, analyzes connections across files, and can process more than 50,000 documents per hour. The company also says Am Law firms already use the platform on active matters, but those throughput and adoption statements remain company-reported rather than independently audited.

What Happened

Newfund Capital led Discernis' $2.5M Seed financing. Triple Impact Capital, Remarkable Ventures, and C2 Ventures also participated, giving the company a syndicate that combines cross-border early-stage investing with New York venture networks and applied-AI exposure.

Discernis said the money will fund further development of its AI and wider adoption among large law firms and corporate legal departments. The announcement did not disclose a valuation, total funding figure, named customers, revenue, or hiring targets, and reviewed primary sources did not provide reliable amounts for earlier financing.

Founder and CEO Robert Kwiatkowski, PhD, brings a technical background that fits the product's hardest sales argument. He earned a doctorate in computer science from Columbia University and worked on AI at AWS before founding Discernis, while the company says his wife's experience with legal discovery helped expose the operating problem behind the product.

Whole-Collection Analysis Changes the Product Promise

Traditional document review reduces a giant collection into a smaller set that people or software can examine. Search terms, technology-assisted review, sampling, ranking, and retrieval can all narrow the work, but the narrowing process also creates a question that matters in litigation: what was left outside the frame?

Discernis Discovery is built around a different promise. Legal teams can provide the review protocol they already use, and the system analyzes documents for responsiveness and relationships across the collection rather than treating every file as an isolated object. Discernis says this lets attorneys surface evidence that appears only when communications, parties, events, and timelines are connected.

That promise raises the standard for the product. A system claiming full-collection analysis has to do more than retrieve an impressive passage; it has to show why a conclusion follows from the record, preserve context across documents, and make the process defensible when another party challenges the result.

Data Control Is Part of the Legal-AI Product

Legal AI carries a deployment problem that ordinary software can sometimes postpone. Privileged communications, investigation records, regulated data, outside-counsel guidelines, and client security requirements can make a third-party model call unacceptable even when the feature works perfectly.

Discernis says its platform can run on-premises or inside a customer's secure cloud, with inference performed within that environment and no call to a third-party commercial AI API. That architecture makes data location and control part of the core product rather than a compliance attachment added after the demonstration.

The same model supports Discernis Investigations, which is designed to map parties, reconstruct timelines, analyze communications, and surface anomalies across internal inquiries and sensitive matters. The commercial opportunity is broader than first-pass discovery if the company can prove that the same architecture transfers across legal workflows without weakening accuracy or explainability.

Why Newfund's Bet Matters

The investor logic rests on technical depth meeting a market with unusually little tolerance for error. Henri Deshays, a General Partner at Newfund, pointed to Kwiatkowski's ability to run advanced AI inside a law firm's walls and his judgment about where that controlled deployment is necessary.

That is a sharper thesis than simply betting on more AI adoption in legal services. Firms already know software can accelerate review; the purchasing decision turns on whether a product can reduce labor and time without creating a new confidentiality, privilege, or defensibility problem that becomes more expensive than the work it replaced.

Discernis reports that its AI reviews more than 50,000 documents per hour and is in use at Am Law firms. Those claims create a credible commercialization story, but the next proof points will be repeatable deployments, independently validated outcomes, customer retention, and evidence that legal teams trust the product across matters with different facts and infrastructure.

What the $2.5M Seed Round Changes

The round gives Discernis room to deepen the product and expand enterprise distribution, two jobs that demand different kinds of discipline. Whole-collection models need technical work around scale, accuracy, explainability, and deployment, while legal sales require security review, procurement patience, implementation support, and credibility with attorneys whose mistakes become part of the record.

Capital can help the company handle both sides of that equation, but it cannot compress institutional trust on command. The defensibility of Discernis' approach will be tested when the collection is ugly, the deadline is bad, the infrastructure is constrained, and the decisive evidence is the document nobody knew to search for.