MIND
MIND provides an AI-powered data loss prevention and insider risk platform for organizations, discovering and classifying sensitive data, detecting risks, remediating issues and preventing leaks across SaaS, AI tools, endpoints, file shares and email. Its mission is to protect organizations’ sensitive data and help them use AI while mitigating data risks.
Facts
- HQ
- Seattle, WA, US
- Sectors
- Computer and Network Security
- Founded
- 2023
- Employees
- 96
- Stage
- Series B
- Founders
- Eran Barak, Itai Schwartz, Hod Bin Noon
Is this your organization? Claim it
Samuel Hill: SANS Survey — 79% of SOCs Use AI but Only 36% Have Defined Workflows, Raising Validation and Accountability Risks
Samuel Hill highlights SANS survey results showing 79% of SOCs use AI or ML but only 36% have integrated it into defined workflows, creating validation, accountability, and staffing gaps that threaten data security despite time-saving benefits.
Seoul court upholds ~US$319M of Korea's record ~US$471M fine against Coupang for 37.5M‑user data breach over missing key management and access controls
A Seoul court upheld South Korea's record roughly $319 million fine against Coupang over a breach affecting about 37.5 million users, citing missing basic safeguards like key management and system access controls; the court kept corrective orders paused briefly while rejecting suspension of the monetary penalty.
MIND named to CRN’s 2026 Stellar Startups list (Security) for its Stress Free DLP platform
MIND was named to CRN’s 2026 Stellar Startups list in the Security category for its Stress Free DLP platform, recognized for AI-powered, autonomous data loss prevention across GenAI, SaaS, endpoints, email and on-prem systems.
CyberFOX survey: 62% of IT/security leaders say staffing shortages leave risks unaddressed; many gaps open 6+ months
CyberFOX’s survey of 361 IT and security leaders finds 62% say limited staffing leaves important security risks unaddressed, with many gaps open six months or more; additional industry surveys report widespread skills shortages and growing workloads from tools like generative AI that increase data exposure.
Samuel Hill: Agentic AI Attacker Accessed Chainalysis Customer Names, Emails and Role Labels — Why ‘Non‑Sensitive’ Data Matters
Samuel Hill examines Chainalysis’s breach in which an attacker using agentic AI accessed customer user IDs, full names, emails, and organization role labels, arguing that so-called non-sensitive identity data can be valuable to attackers and should be re-evaluated by defenders.
Samuel Hill highlights survey: 58% of enterprises run 11+ data security tools and only 7% have unified stacks
Samuel Hill highlights a 2026 survey of 1,064 practitioners showing 58% of large organizations run 11+ data security tools while only 7% have a unified stack, causing manual correlation, inconsistent policies, and visibility gaps.
Samuel Hill: Uber issues enterprise ChatGPT to 30,000 employees — licensing is necessary but insufficient
Samuel Hill applauds Uber for rolling enterprise ChatGPT to 30,000 employees but warns that licensing is only the first step: companies must enforce separation between sanctioned AI and personal accounts, block unapproved services for sensitive data, and set clear policies so security teams can audit and control data use.
Landen Brown warns PHP2Shell RCEs (CVE-2026-109618, CVE-2026-109619) actively exploited — unpatched PHP servers risk data, creds, and API keys
Landen Brown explains that two critical unauthenticated remote code execution flaws in PHP (CVE-2026-109618 and CVE-2026-109619) are being actively exploited, meaning unpatched PHP servers—used by ~70% of sites and platforms like WordPress—could expose customer data, credentials, databases, and API keys; operators must verify patches and assess what compromised servers could reach.
Samuel Hill: N-able survey finds 75.5% of security teams won’t trust AI remediation without human review
N-able’s survey of 595 security pros shows 75.5% won’t trust AI remediation without human review, citing gaps in context and control—poorly explained alerts, duplicate signals, and lack of human overrides undermine confidence in autonomous DLP.
Samuel Hill: Stolen login caused Meta personal data to land in GitHub, exposing policy gaps
Samuel Hill explains that a stolen login led to Meta employees' personal data being pushed into a GitHub repository, arguing that company data policies break down when discovery of sensitive information in code repos is missing.
Samuel Hill: Identity Alone Can’t Secure AI Agent Access; Delinea Finds Firms Suspect Agents Reached Extra Data
Samuel Hill explains that relying solely on agent identity is insufficient to ensure data security, noting Delinea’s finding that many firms suspect AI agents accessed data beyond their intended tasks and outlining what additional controls are needed for comprehensive protection.
Samuel Hill: Simple DLP Reduces Workarounds After Zapier Finds 80% of Executives Report Staff Bypass AI Policies
Samuel Hill explains that security friction in AI policies leads employees to bypass rules, and argues that simple data loss prevention (DLP) tools can reduce workarounds and keep data safer.
Samuel Hill reports AI pentest agent enabled breaches of South Korean banks by exploiting side systems, increasing risk to all data copies
Samuel Hill explains how an AI penetration-testing agent enabled breaches of South Korean banks via peripheral systems, arguing that AI-driven attackers worsen data sprawl and outlining approaches to reduce that risk.
Samuel Hill: AI-Related Alerts Rose 685% in Four Months, 94.1% Are Noise — Autonomous DLP Must Read Content and Context
Samuel Hill reports that AI-related alerts rose 685% in four months but 94.1% are false positives, arguing that autonomous DLP must analyze content and context to surface real data risk.
Samuel Hill reports Pentagon DMDC breach exposed unencrypted SSNs on file share for nine months
Samuel Hill reports that the Pentagon DMDC breach left unencrypted Social Security numbers accessible on a file share for nine months, and recommends checks CISOs should perform on their own file shares to prevent similar exposure.
Hod Bin Noon announces MIND integrates Claude Compliance API to add context-aware classification for Claude Enterprise and Platform
Hod Bin Noon explains that MIND integrates with the Claude Compliance API to add context-aware classification to Claude Enterprise and Claude Platform activity, enhancing data loss prevention (DLP). The post outlines how the integration works and what it enables for smarter DLP.
Gyazo data breach exposed OCR text and location metadata from 490 million screenshots up to seven years old
A breach at screenshot service Gyazo exposed OCR-processed text and location metadata from 490 million images, including files up to seven years old, highlighting risks from long-retained data in everyday tools.
Samuel Hill: $8,000 AI-Orchestrated Skimmer Breached 100 Companies — Why DLP Failed and What Could Stop It
Samuel Hill examines a $8,000 AI-orchestrated payment skimmer that compromised 100 companies, explains why typical DLP programs would not have prevented it, and outlines what defenses would have been effective.
Samuel Hill: EY Finds 98% of Large Firms Have AI Policies but 25% Cannot Detect Rogue Agents, Raising Data Security Alarm
Samuel Hill highlights that while 98% of large companies report having AI policies, about 25% cannot detect rogue AI agents, arguing that full data security — not additional policy — is needed to close this governance gap.
Anthropic VP Dario Amodei Urges Frontier AI Pause and Data-First Protections After OpenAI–Hugging Face Exposure
Dario Amodei urges leading AI labs to slow development and prioritize protecting data at its source, arguing recent incidents like OpenAI's Hugging Face exposure show safeguards must be applied to the data itself before agents access it.
Samuel Hill: ShinyHunters Claims McKesson Salesforce Leak of Support Cases and Patient Details, Highlighting Free‑Text Data Risks
Samuel Hill examines a ShinyHunters claim that they obtained McKesson support cases and patient details from the company's Salesforce, arguing that unstructured free-text data poses the greatest exposure risk.
Samuel Hill announces the MIND Claude Connector: read-only Claude access to MIND over MCP for issues, policies, and inventory
Samuel Hill explains a read-only connector that links Claude to MIND over MCP, enabling plain-English queries about issues, policies, and inventory with quick, minutes-long setup and IT install steps included.
Samuel Hill reports Truffle Security: 88% of 10,616 leaked AWS keys still authenticate, some active four years
Samuel Hill reports that Truffle Security’s analysis found 88% of 10,616 leaked AWS keys still authenticate — some active up to four years — underscoring that credentials rarely get rotated. The piece argues the real fix is preventing key leakage rather than relying on rotation to mitigate exposure.
Samuel Hill explains how to distinguish AI-native data security from AI-washed AI SOC agents
Samuel Hill explains how to distinguish genuinely AI-native data security solutions from AI-washed marketing claims, giving buyers practical guidance to evaluate AI SOC agents amid heavy hype.
Samuel Hill: Research Shows AI Agents Access Twice the Authorized Data, Undermining Human-First Security
Samuel Hill highlights research finding that AI agents access twice as much data as humans are authorized for and argues that human-first security models cannot adequately control agent behavior, calling for a new approach to complete data security.
People
See all 5-
Eran Barak Co-Founder & CEO
-
Hod Bin Noon Co-Founder & VP of R&D
-
Eran Barak Co-Founder Co-Founder & CEO