Skip to content

Bonfy AI

AI-native platform that classifies and protects unstructured data as it moves through email, SaaS apps, copilots, and AI agents.

Visit Website ↗ + Add to Compare
63/100Incremental Innovator

Overview

Bonfy AI builds Bonfy Adaptive Content Security (ACS), a platform that unifies data-in-motion, data-at-rest, and data-in-use protection across email, files, SaaS applications, collaboration tools, and generative-AI workflows. Its entity-aware classification engine is designed to understand the people, customers, and context behind content, rather than relying on static keyword or regex rules, and it extends into agent-level controls through an MCP server that lets AI agents check content safety mid-reasoning.

The company was founded by security veterans Gidi Cohen and Danny Kibel and raised $9.5 million to build out its Adaptive Content Security platform, which won Best AI Security Solution at The Hacker News Cybersecurity Stars Awards 2026. Bonfy targets the emerging problem of shadow AI and agentic workflows moving sensitive data outside traditional DLP boundaries, integrating with Microsoft 365, Google Workspace, Salesforce, HubSpot, and Slack.

Innovation Matrix Assessment

Innovation Velocity 7/10

Shipped an MCP-based agent-safety layer and ACS 2.0 within roughly a year of a $9.5M raise, moving fast into the shadow-AI/agentic-data problem.

Operational Value 7/10

Consolidates DLP-style controls across email, SaaS, and AI agents into one classification engine, reducing tool sprawl for data-security teams.

Market Momentum 5/10

Early-stage funding ($9.5M) and one industry award; customer names and deployment scale are not yet public.

Category Disruption 6/10

Agent-level, in-reasoning data checks are a genuinely newer approach to DLP, though the space is crowding fast with GenAI security entrants.

Real-World Efficacy 5/10

Efficacy claims are vendor-published; no independent test results or named breach-prevention case studies were found.

Enduring Relevance 8/10

Shadow AI and agentic data leakage are accelerating problems that will matter more, not less, over the next several years.

Why CISOs Should Care

Gives CISOs a single control plane for data leaving the org through GenAI tools and agents, a blind spot most legacy DLP was never built for.

What Makes It Different

Entity-aware classification plus an MCP server that lets AI agents themselves query content safety mid-task, rather than scanning after the fact.

The Matrix Verdict

63/100 — INCREMENTAL INNOVATOR

A promising, well-funded entrant addressing a real and growing gap in AI-era data protection, but still early on independent validation.

Editorial Note: Claims vs. Verified Findings

Performance and accuracy figures come from company and award materials; no third-party efficacy testing was located.

Sources