Skip to content
Feterand
Service · AI security

Your AI, hardened against attack.

We secure AI applications and the infrastructure behind them: prompt injection, data leakage, agent permissions and exposed models. Ready for the EU AI Act, ISO 27001 and GDPR.

Bringing an LLM into your company opens a new attack surface: instructions hidden in documents, agents with more permissions than they should have, sensitive data that ends up in a prompt or with an outside provider.

We attack your AI the way an adversary would, close the gaps in order of risk, and stay on watch — the model, the agents and the systems they run on.

What's included

  • LLM and agent red teaming: prompt injection, jailbreaks, exfiltration.
  • Data control: what the model sees, what leaves and where it goes.
  • Least privilege and audit trails for agents and tools.
  • Hardening and 24/7 monitoring of the infrastructure serving the models.
  • Compliance: EU AI Act, ISO 27001, GDPR.

How we work

  1. 01

    AI audit

    Red teaming of models and agents, review of data flows, permissions and infrastructure.

  2. 02

    Remediation

    A risk-prioritised plan and hands-on help fixing it.

  3. 03

    Monitoring

    Continuous monitoring, real-time alerts and incident response.

What you get

  • AI red-teaming report, every finding prioritised by risk.
  • Remediation plan with owners and order of execution.
  • Data and permissions map: what each model and agent can see.
  • Supporting documentation for the EU AI Act, ISO 27001 and GDPR.

Plans and pricing

AI audit

Fixed project, 1–3 weeks

Custom

You're launching (or already run) AI in production and want to know where you're exposed.

  • Red teaming of models and agents
  • Review of data, permissions and infrastructure
  • Risk-prioritised report
  • Results session with leadership
Request a proposal
Recommended

Ongoing security

Monthly fee, min. 6 months

Custom

Your AI changes every week and security has to keep pace.

  • Everything in AI audit
  • Hands-on remediation
  • 24/7 monitoring and alerts
  • Recurring red teaming after every significant change
  • AI Act / ISO 27001 / GDPR documentation kept current
Request a proposal

Incident response

Hour bank or per incident

Custom

Something has happened (or you fear it will) and you need someone to act now.

  • Incident containment and analysis
  • Closing the breach and hardening
  • Incident report
  • Recommendations so it doesn't happen again
Request a proposal

Every project is closed with a written proposal: scope, timeline and price before we start.

They trust us

  • SDM · Sistemas Dosificación Medicamentos

Case studies

SDM · Sistemas Dosificación Medicamentos · Healthcare · medication dosing

We recovered and modernised a medication-dosing management system

incidents
−80 %
incidents
Challenge
A medication-dosing management system critical to daily operations, built on outdated legacy code, with open security issues and no room to evolve.
Solution
We recovered the legacy code, brought it up to date and patched every security issue. On that sound base we built new features and added new functions.
Result
An up-to-date, secure system ready to keep growing, without losing what already worked.

Shall we talk about your case?

We reply with a first assessment of scope and price.