Artificial Intelligence

Knostic Secures $11 Million to Rein in Enterprise AI Data Leakage, Oversharing 

Knostic provides a “need-to-know” filter on the answers generated by enterprise large language models (LLM) tools.

Knostic provides a “need-to-know” filter on the answers generated by enterprise large language models (LLM) tools.

Knostic, a Virginia startup building technology to manage data leakage and oversharing with enterprise-class AI tools, has banked $11 million in a new funding round. 

The company, which emerged from stealth last April with ambitious plans to provide a “need-to-know” filter on the answers generated by large language models, said the seed-stage financing was led by Bight Pixel Capital.

Silicon Valley CISO Investments) and previous backers DNX Ventures and Seedcamp also took equity positions.

Knostic, founded by veteran security pros Gadi Evron and Sounil Yu,  has raised a total of $14 million to build and market a “knowledge control layer” that ensures that AI outputs adhere to a company’s need-to-know principles.

The Knostic approach is simple: instead of blocking information outright, the product ensures that AI systems only share data that a user is cleared to see. For example, an intern should not be able to receive answers on sensitive financial matters or data that contains intellectual property. 

“The problem is that these tools just can’t keep a secret,” Evron explained, noting that AI-powered products like Microsoft 365 Copilot and Glean lack the ability to discern what’s appropriate and in what context.

Advertisement. Scroll to continue reading.

Knostic and its investors are betting there’s a big market for technology that integrates directly with AI systems to control the flow of information and applies need-to-know filters that adjusts AI responses based on a user’s clearance. 

The Knostic product line also handles continuous oversharing detection by regularly simulating queries to catch potential data leaks.

Knostic is also providing tools for contextual curation to deliver sanitized answers that remain useful without exposing confidential details and real-time alerts to help administrators to quickly tighten permissions when needed.

The company fits into the emerging category of AI-driven data loss prevention (DLP) that tackle oversharing with LLMs and AI-powered productivity tools. Knostic’s technology bridges this gap by integrating principles of identity and access management with AI governance, promising tooling to ensure that sensitive information is shared on a need-to-know basis.

Related: Knostic Emerges From Stealth With Enterprise Gen-AI Access Controls

Related: Enterprise AI Security Firm TrojAI Raises $5.75M in Seed Funding

Related: Generative AI Startup Nexusflow Raises $10.6 Million

Related: Harmonic Lands $7M Funding to Secure Generative AI Deployments

Related Content

Cybersecurity Funding

The startup brings AI to data loss prevention to provide visibility into intent, context, and risk.

Funding/M&A

Proofpoint said Hornetsecurity brings in nearly $200 million in annual recurring revenue, with a 20% year-over-year growth rate.

Artificial Intelligence

Agentic AI speeds operations, but requires clear goals, least privilege, auditability, red‑teaming, and human oversight to manage opacity, misalignment, and misuse.

Email Security

Hackers are abusing the Microsoft 365 Direct Send feature to deliver phishing emails that bypass email security controls.

Artificial Intelligence

Microsoft recently patched CVE-2025-32711, a vulnerability that could have been used for zero-click attacks to steal data from Copilot.

Cybersecurity Funding

Data security firm MIND has raised $30 million in Series A funding to expand its R&D and go-to-market teams.

M&A Tracker

Enterprise cybersecurity giant Proofpoint is buying Germany-based Microsoft 365 security solutions provider Hornetsecurity.

Cloud Security

Orca positioned the deal as an expansion of its capabilities into the realm of AI-based autonomous remediation and prevention. 

Copyright © 2026 SecurityWeek ®, a Wired Business Media Publication. All Rights Reserved.

Exit mobile version