Skip to main content

Verno Labs

Category: AI in Cybersecurity

Offensive security lab for AI agents—custom offensive LLMs that adversarially test agentic systems, then drive detection, enforcement, and remediation from findings. Verno Labs was founded in 2025. The company is led by Pavlos Mitsoulis. Based in London, United Kingdom. Team size: 2-10. Total funding raised: €1.36M. Latest round: Pre-Seed. Key investors include Marathon Venture Capital, Executives from AWS, Executives from Meta, Executives from Google, Executives from Hack The Box, Executives from Workable.

Founded
2025
Headquarters
London, United Kingdom
Team size
2-10
Total funding
€1.36M

Value proposition

Closed loop of autonomous red teaming (Verno Red), real-time runtime detection (Verno Guard), and auto-generated remediation (Verno Fix) so enterprises find AI-agent vulnerabilities before attackers and harden continuously.

Products and solutions

Verno Red (autonomous adversarial attack campaigns), Verno Guard (real-time runtime detection across agents and tool calls), Verno Fix (auto-generated remediations from findings), VernoBench / proprietary 150+ technique attack taxonomy

Unique value

Defense is downstream of offense: custom offensive LLMs and continuous autonomous red teaming produce the attack intelligence that powers runtime detection and remediation—vs. guardrail-only or prompt-injection-only peers.

Target customer

Enterprise security teams (CISOs), AI engineering departments, and organizations deploying autonomous agentic workflows.

Industries served

Cybersecurity, Enterprise Software, Financial Services, E-commerce, Healthcare

Technology advantage

Leverages the founders' deep expertise in MLOps and large-scale ML infrastructure (from leadership roles at On, King, and Expedia) to build a low-latency, scalable security interceptor that monitors agent-tool interactions in real-time.

How they differentiate

Verno Labs differentiates with offense-first AI red teaming (custom offensive LLMs, multimodal text/voice coverage, proprietary attack taxonomy) that continuously feeds Guard/Fix—vs. competitors focused mainly on prompt injection, model integrity, or static guardrails.

Main competitors

Lakera, Protect AI, Prompt Security, HiddenLayer

Key partnerships

Marathon Venture Capital (Lead Investor), Strategic Angel Network (Executives from AWS, Meta, Google, Hack The Box, and Workable)

Notable customers

Blueground (Stratos Pavlakis, CTO), Nofire / NoFire AI (Panos Moustafellos, CTO)

Major milestones

Incorporated as PALLMA AI LIMITED in London (Nov 2025), Exited stealth / announced €1.36M ($1.6M) Pre-Seed led by Marathon VC (Jan 2026), Rebranded to Verno Labs; launched Verno Red/Guard/Fix offensive security platform (Jun 2026), Exhibiting at Black Hat USA 2026 (booth 6103)

Growth metrics

Emerged from stealth Jan 2026 with $1.6M Pre-Seed; rebranded to Verno Labs Jun 2026; ~7 employees (LinkedIn); reported ~1260% website-visit growth in Tip-Offs AI agent security ranking.

Market positioning

Early-stage offensive AI security lab (AISec) specializing in continuous red teaming and runtime protection for enterprise agentic systems; targeting CISOs and AI engineering teams.

Geographic focus

United Kingdom, European Union, and North America.

Patents and IP

No registered patents disclosed as of latest update

About Pavlos Mitsoulis

Pavlos Mitsoulis is the Co-Founder and CEO of Pallma. He previously served as the Director of Data, Machine Learning, and AI at On (On-Running), where he led the global data and AI strategy. His prior experience includes leadership roles as a Machine Learning and Engineering Leader at King (Activision Blizzard) and a Staff Data Scientist at Expedia Group. He is a recognized AWS Machine Learning Hero and the co-creator of Sagify and Kenza.

Latest news about Verno Labs

More AI in Cybersecurity companies

Official website: