02 · Cybersecurity & Identity

AI Application Security Assessment

Authorised testing of chatbots, RAG and agents for AI-specific risks.

Pricing is provided after a short scoping conversation, based on your requirements.

Request a quote

The problem

AI features introduce new attack paths — prompt injection, data leakage and over-permissioned tools.

A scoped, authorised assessment of your AI application covering prompt injection, access boundaries, data leakage, and the permissions of connected tools.

What's included

  1. 01 Threat model of the AI application
  2. 02 Authorised testing within agreed scope
  3. 03 Findings on injection, leakage and tool permissions
  4. 04 Risk-ranked remediation guidance
  5. 05 Retest of agreed fixes if scoped

Not included

  • Testing systems you are not authorised to approve
  • Model provider infrastructure
  • Ongoing monitoring

How we work

  1. 01 · Discovery

    Agree written scope, test accounts and rules of engagement.

  2. 02 · Delivery

    Threat-model, test and validate findings with your team.

  3. 03 · Handover

    Documentation, a walkthrough for your team and an agreed list of recommended next steps.

Related services