Fabraix GTM brain

Fabraix sells Nyx, an automated security testing agent for organizations deploying customer-facing AI agents.

category

AI agent security testing

market profile

  • text: Customer-facing agents change with their models, permissions and tools, creating a need for repeated security testing.
  • text: AI security buyers can choose automated red-team platforms, open-source test frameworks or traditional security engagements.

personas

  • title: Head of application security
    pains:
    • A one-off red team becomes stale as agent behavior and integrations change.
    • Needs reproducible findings that engineering teams can validate and fix.
  • title: AI product engineering leader
    pains:
    • Customer-facing agents can fail in ways ordinary application testing does not reveal.
    • Release cycles outpace manually scheduled adversarial testing.
  • title: CISO at a large enterprise
    pains:
    • Needs evidence to support release decisions for AI handling customer interactions.
    • Must manage testing expense across a growing estate of agents.

icp

  • text: Enterprises shipping customer-facing AI agents, including large companies with security review requirements.
  • text: AI product teams whose agents change frequently and interact with external systems or content; no fixed headcount threshold is published.

message pillars

  • text: Continuously test customer-facing AI for security weaknesses.
  • text: Connect findings to engineering remediation.
  • text: Build trust through adversarial testing.

objections

  • objection: Will this expose internal systems or require source access?
    response: The vendor describes Nyx as black-box; confirm target access, permitted scope and data handling for the planned test.
  • objection: How is it better than our existing open-source red team framework?
    response: Run both on an authorized representative agent and compare useful validated findings, coverage and engineering effort.

competitors

  • name: Promptfoo
    why: Offers developer-oriented AI evaluations and red teaming.
  • name: Mindgard
    why: Commercial platform for testing and red teaming AI systems.
  • name: Giskard
    why: Offers AI testing and continuous agent red teaming.

keywords

  • AI agent red teaming
  • automated AI security testing
  • AI agent vulnerability assessment
  • LLM security testing
  • agent security evaluation
  • continuous AI red teaming

voice

  • Technical and adversarial
  • Confident and direct
  • Security practitioner focused

signals

  • text: A company launches a customer-facing agent or grants it new integrations and permissions.
  • text: An application security team is establishing repeatable AI security testing before releases.

hiring

  • AI researchers
  • Product engineers
  • Forward-deployed engineers

pricing

model: unknown

motion

sales_led

channels

  • YC launch
  • LinkedIn
  • public playground

confidence

medium

Founders

  • name: Ahmed Aly
    title: Founder/CEO
    role: CEO
    bio: Co-founder @ Fabraix, building a frontier hacker agent that finds failure modes and security exploits in AI systems before real users do. Previously the first data scientist at Sequoia-backed Two, where he built AI systems for adversarial environments from scratch and scaled them to process $1B+ in annual B2B transactions, averting $50M in losses for multiple Fortune 150 companies. A published researcher with multiple papers in Q1 journals and a UCL PhD drop-out.
  • name: Ibrahim Abdu
    title: Founder/CTO
    role: CTO
    bio: Red-teamed AI agents in Meta's Privacy org; built an SRE agent triaging 3M+ errors/month, saving 10,000+ engineering hours a year. Built a compiler for insider-trading detection at TradingHub: $100M+ in daily trades; Oxford PPE, top 8% of cohort.

Back to YC companies