Nightfall

Nightfall AI is a data loss prevention (DLP) and insider risk management platform that detects, classifies, and controls the movement of sensitive data across SaaS applications, endpoints, browsers, email, and generative AI tools. The platform detects categories such as PII, PHI, PCI, secrets and credentials, and uses LLM-based file classifiers to identify document types including financial statements, internal source code, HR records, contracts and NDAs, product roadmaps, and tax and audit documents. Remediation actions include blocking, redaction, quarantine, deletion, encryption, permission revocation, and prompting employees for business justification ("employee coaching").

Product areas described by the company include data exfiltration prevention (including shadow AI prevention and data lineage reconstruction), AI agent security (interception of AI agent tool calls, prompt-injection blocking, Model Context Protocol server discovery and policy enforcement), data detection and response, posture management, data encryption for email, historical data discovery and classification, and developer APIs for filtering sensitive data from GenAI apps and datasets. Documented integrations cover Slack, GitHub, Google Drive, Gmail, Confluence, Jira, Microsoft 365 (OneDrive, Teams, Exchange Online, SharePoint), Salesforce, Zendesk, Notion, and Anthropic Claude Enterprise via the Claude Compliance API, alongside endpoint agents for macOS and Windows and Chrome and Firefox browser extensions.

The company positions the platform against legacy DLP, which it characterizes as reliant on pattern matching with 5-25% accuracy and lengthy tuning cycles, versus Nightfall's stated 95% out-of-the-box classification accuracy from 100+ AI-based models, LLM file classifiers, and computer vision models.

Founding story

Co-founder and CEO Rohan Sathe led backend teams at Uber Eats, building applied ML services such as ETA prediction and supply/demand forecasting, where he worked with petabyte-scale data spread across many systems. He cites that experience plus industry lessons from incidents such as Uber's 2016 breach — in which exposed credentials in code on GitHub were used to reach AWS data — as showing how data sprawl, credentials, and cloud infrastructure combine to create outsized risk. Co-founder Isaac Madan was previously an investor at Venrock focused on SaaS, AI, and security. The pair founded the company in 2018 on the premise that machine learning could make DLP more accurate and easier to deploy than legacy tools.

Business model

Nightfall sells a B2B SaaS data security platform to enterprises, deployed through API-based SaaS integrations, endpoint agents, and browser plugins, complemented by a developer platform exposing detection as cloud-hosted APIs and SDKs. The sources do not disclose pricing or contract structure.

Not disclosed in the available sources beyond the enterprise software/platform model; the company markets a demo-led sales motion and an ROI estimator.

Traction

Named customers and references include Snyk, Unit21, Neural Payments, Telnyx, UserTesting, and Rightway. At the time of the 2022 Series B the company reported over 4,000% growth in data volume scanned since 2021 and more than 40 million sensitive data findings remediated for customers, and said it planned to more than double headcount over the following year. Current marketing cites 95% detection precision, 20x average ROI, and 80% automated remediation. Total disclosed funding is $60M across two rounds.

Latest developments

Recent announcements include Forensic Search and App Intelligence for insider risk visibility (February 2026), three new detectors for image-based PII and regional coverage (March 2026), and a live integration with Anthropic's Claude Compliance API (June 2026). Product messaging has shifted toward AI agent and MCP security, including MCP server discovery and inventory, hook-level interception of agent tool calls, prompt-injection blocking, and tool-call governance, supported by the State of Agentic Data Security 2026 report and the 2026 AI Agent Risk & Action Report.