Rig Security Raises $12M to Track AI Agents

Key Takeaways

• Rig Security raised $12 million in seed funding led by Ten Eleven Ventures and Brightmind Partners to track AI agents operating under employee accounts
• Rig's RICE identity correlation engine achieves over 96% accuracy in distinguishing AI agent activity from human user sessions
• America.gov launched on September 29, 2026, using Grok and Gemini to help Americans access government information from 29,000 websites through a single search
• StackHawk's Wingman works with Claude Code, Cursor, GitHub Copilot, Codex, and Antigravity to fix AI coding security vulnerabilities at $10 per user per month
• Komprise launched Universal File MCP to reduce AI agent data bloat by providing a single interface for accessing enterprise unstructured data
• Rig Security's Bifrost endpoint sensor can block risky AI agent actions in real time while allowing employees to continue working
• Fortune 200 companies in financial services, insurance, health care, and technology are already using Rig Security's platform
• America.gov was built by the General Services Administration and will eventually support passport applications, Medicare enrollment, and government job applications by 2027
• Early users of StackHawk Wingman fixed over 7,500 vulnerabilities with 98 percent of fixes remaining resolved
• Rig Security was founded by CEO Guy Kozliner, a Wiz alumnus, and includes Wiz co-founder Ami Luttwak among its investors

Rig Security launches with $12M to track AI agents using employee accounts

Rig Security, an Israeli startup, launched with a $12 million seed round to help companies track AI agents that operate under employees' existing accounts. Its platform uses the Rig Identity Correlation Engine, or RICE, to match identities across systems with over 96% accuracy. A lightweight endpoint sensor called Bifrost can separate an agent's activity from a user's and block risky actions in real time. Fortune 200 companies in financial services, insurance, and health care already use Rig, which is available through the AWS Marketplace and CrowdStrike Marketplace.

An employee or AI? A $12M Israeli startup wants to tell the difference

Rig Security raised $12 million to help companies tell whether an action was performed by an employee or an AI agent. The startup's platform maps identities across cloud systems, networks, and identity providers to determine what is actually using those identities. It can distinguish an AI agent's session from a human user's session and allow the employee to keep working while blocking risky agent actions. Rig says its customers include Fortune 200 companies in financial services, insurance, health care, and technology.

Wiz veteran raises $12 million Seed to secure AI agents at Rig Security

Rig Security, founded in early 2025 by CEO Guy Kozliner, a Wiz alumnus, raised a $12 million seed round led by Ten Eleven Ventures and Brightmind Partners. The CrowdStrike Falcon Fund also participated, along with cybersecurity founders including Wiz co-founder Ami Luttwak. The company's platform includes RICE, an identity correlation engine with over 96% accuracy, and Bifrost, an endpoint sensor that can enforce policies on AI agent activity. Rig has 20 employees and is used in production by Fortune 200 organizations.

Five things to know about Trump's new AI website

President Trump launched America.gov, a website that uses AI to help Americans find government information from 29,000 websites through a single search. The site uses Grok and Gemini and was built by the General Services Administration with the National Design Studio. Trump said the site cannot be hacked and does not collect personal information. It will eventually allow users to apply for passports, enroll in Medicare, update legal names, and apply for government jobs in 2027. The site is free, ad-free, and does not provide campaign or party information.

Everything to know about America.gov, Trump's new AI website

President Trump launched America.gov on September 29, 2026, as a central hub for government information and services. The website features a search bar, popular government services, and a frequently asked questions section. Trump promoted the site on social media and said it would make government services easier to access. Critics raised concerns about transparency, accountability, functionality, and user experience. The site's development and launch details have remained largely unclear.

AutoEnergy improves energy forecasting with automated feature engineering

Researchers developed AutoEnergy, an automated feature engineering tool for predicting energy use. It works with AutoML to build forecasting models across 18 real-world energy datasets. AutoEnergy cut forecasting error by 19.52% to 84.72% compared with baseline methods. It also ran 1.31 to 4.41 times faster. When paired with Decision-Focused Learning for battery storage, it reduced operating costs by 22.9% to 56.5%.

SeLMRoute uses semantic evidence to improve AI model routing

SeLMRoute is a new framework for routing queries to the best large language model. It extracts interpretable semantic evidence from queries, such as reasoning needs and external knowledge use. A supervised router uses this evidence to estimate model performance. On LLMRouterBench with 11,481 queries and 20 models, it achieved 72.08% average accuracy. It also improved performance by a mean PerfGain of 2.66% in cost-aware settings.

AI investment risks grow as technology reshapes the economy

The article warns that everyday people are now effectively AI investors as the technology spreads through the economy. The author says AI-driven investing has created new risks that many are not prepared for. The U.S. stock investor culture over the past 40 years has already fueled speculation. The rise of AI makes these dangers worse. The article calls for greater awareness and steps to reduce AI-related financial risks.

HPE links AI sovereignty with quantum security for data protection

HPE Labs is connecting AI sovereignty with quantum security to protect sensitive data. Director Andrew Wheeler says customers regularly ask about AI, quantum, and data sovereignty. Over 170 countries have data protection rules, and AI is increasing the urgency. HPE is exploring quantum key distribution and confidential computing for AI. The company also took part in Project Glasswing to strengthen security design and protect data in transit.

Komprise launches universal interface to reduce AI agent data bloat

Komprise introduced Universal File MCP to address MCP bloat caused by too many AI tool connections. The tool gives AI agents a single interface to access enterprise unstructured data across multiple storage types. It filters data to reduce token costs and improve accuracy. It enforces user permissions and uses the Komprise Global Metadatabase for consistent schemas. Analysts say it simplifies how AI accesses large-scale unstructured data.

Finastra Uses AI to Help Banks Fix Payment Processing Issues

Finastra launched a new Repair Recommendations capability that uses AI to identify payment discrepancies and suggest corrections. The tool reduces time and effort needed to resolve payment exceptions and lowers error rates. It helps banks modernize payment hubs that route payments based on cost, speed or priority. Human review is still required for all recommended actions.

StackHawk Launches Wingman to Fix AI Coding Security Flaws

StackHawk launched Wingman, a tool that finds and fixes security vulnerabilities during AI-assisted coding sessions. It works with Claude Code, Cursor, GitHub Copilot, Codex and Antigravity. Early users fixed over 7,500 vulnerabilities with 98 percent of fixes staying resolved. Wingman costs $10 per user per month and includes 50 scans per user each month.

FinAutoRubric Generates Rubrics for Evaluating Financial Research Agents

FinAutoRubric is a method that uses expert guidance to automatically create evaluation rubrics for financial research agents. Experts set reusable rules while agents handle task-specific rubric generation and validation. The system was tested on three expert-authored finance benchmarks and matched human scoring. The team also released a 100-query benchmark covering 78 tasks across eight asset classes.

Sources

• Rig Security launches with $12M to watch AI agents that run under employees' accounts
• An employee or AI? A $12M Israeli startup wants to tell the difference
• Wiz veteran raises $12 million Seed to secure AI agents at Rig Security
• Five things to know about Trump’s new AI website
• Everything to know about America.gov, Trump's new AI website
• Automated feature engineering, AutoML, and decision-focused learning for improved energy consumption forecasting
• SeLMRoute: Probabilistic Semantic Evidence for Large Language Model Routing
• We’re all AI investors now, and that’s risky
• Data sovereignty rises as AI and quantum reshape security
• Komprise combats 'MCP bloat' with a universal interface for AI agents to access enterprise data
• Finastra Deploys AI to Help Banks Tackle Payment Processing Issues
• StackHawk launches Wingman to fix AI coding flaws
• FinAutoRubric: Expert-Guided Automatic Rubric Generation for Evaluating Financial Research Agents