AI Security Assessment

Many organisations lack clear visibility into where AI is being used, what it can access and how effectively it is governed.

This assessment has been designed to help you quickly understand your current AI security maturity, identify potential areas of exposure and highlight opportunities to strengthen your security posture.

Step 01 / 04

Identity

AI introduces a rapidly growing number of non-human identities, including AI agents, service accounts, APIs and automated workloads. Understanding who, or what, has access is becoming critical to maintaining security.

Do you know which AI applications, agents and automated services currently have access to your environment?

[ ] Controlled – Clear visibility with consistent governance and security controls. [ ] Partially Controlled – Some controls exist but are inconsistent, incomplete or difficult to enforce. [ ] Uncontrolled – Limited visibility, governance or confidence in this area.

Are you confident AI systems only have access to the applications and data they genuinely need?

[ ] Controlled – Clear visibility with consistent governance and security controls. [ ] Partially Controlled – Some controls exist but are inconsistent, incomplete or difficult to enforce. [ ] Uncontrolled – Limited visibility, governance or confidence in this area.

Are access policies applied consistently across employees, third parties and AI-driven identities?

[ ] Controlled – Clear visibility with consistent governance and security controls. [ ] Partially Controlled – Some controls exist but are inconsistent, incomplete or difficult to enforce. [ ] Uncontrolled – Limited visibility, governance or confidence in this area.

If an AI system performed an unexpected action, could you quickly identify exactly what happened?

[ ] Controlled – Clear visibility with consistent governance and security controls. [ ] Partially Controlled – Some controls exist but are inconsistent, incomplete or difficult to enforce. [ ] Uncontrolled – Limited visibility, governance or confidence in this area.

Step 02 / 04

Network

Modern AI services require secure access to applications and data regardless of where users, devices or AI services are located. Traditional perimeter security is no longer sufficient.

Is access to applications based on identity and trust rather than network location?

[ ] Controlled – Clear visibility with consistent governance and security controls. [ ] Partially Controlled – Some controls exist but are inconsistent, incomplete or difficult to enforce. [ ] Uncontrolled – Limited visibility, governance or confidence in this area.

Can AI applications securely access business systems without relying on broad network connectivity or VPN access?

[ ] Controlled – Clear visibility with consistent governance and security controls. [ ] Partially Controlled – Some controls exist but are inconsistent, incomplete or difficult to enforce. [ ] Uncontrolled – Limited visibility, governance or confidence in this area.

Is access automatically adjusted based on user identity, device posture and risk?

[ ] Controlled – Clear visibility with consistent governance and security controls. [ ] Partially Controlled – Some controls exist but are inconsistent, incomplete or difficult to enforce. [ ] Uncontrolled – Limited visibility, governance or confidence in this area.

Are security policies enforced consistently across office, remote and cloud environments?

[ ] Controlled – Clear visibility with consistent governance and security controls. [ ] Partially Controlled – Some controls exist but are inconsistent, incomplete or difficult to enforce. [ ] Uncontrolled – Limited visibility, governance or confidence in this area.

Step 03 / 04

Cloud

AI is increasing the speed at which cloud services are deployed, creating new risks if organisations cannot maintain visibility and prioritise genuine security issues.

Do you have visibility of all AI-enabled cloud services and workloads across your environment?

[ ] Controlled – Clear visibility with consistent governance and security controls. [ ] Partially Controlled – Some controls exist but are inconsistent, incomplete or difficult to enforce. [ ] Uncontrolled – Limited visibility, governance or confidence in this area.

Can your security team quickly identify the AI-related risks that matter most

[ ] Controlled – Clear visibility with consistent governance and security controls. [ ] Partially Controlled – Some controls exist but are inconsistent, incomplete or difficult to enforce. [ ] Uncontrolled – Limited visibility, governance or confidence in this area.

Do different teams share a consistent view of cloud security risk and ownership?

[ ] Controlled – Clear visibility with consistent governance and security controls. [ ] Partially Controlled – Some controls exist but are inconsistent, incomplete or difficult to enforce. [ ] Uncontrolled – Limited visibility, governance or confidence in this area.

Are you confident new AI services can be introduced without creating unknown security exposures?

[ ] Controlled – Clear visibility with consistent governance and security controls. [ ] Partially Controlled – Some controls exist but are inconsistent, incomplete or difficult to enforce. [ ] Uncontrolled – Limited visibility, governance or confidence in this area.

Step 04 / 04

Data

AI systems are only as powerful as the data they can access. Without appropriate governance, sensitive information can easily become exposed, shared or misused.

Do you know what sensitive data employees and AI systems can currently access?

[ ] Controlled – Clear visibility with consistent governance and security controls. [ ] Partially Controlled – Some controls exist but are inconsistent, incomplete or difficult to enforce. [ ] Uncontrolled – Limited visibility, governance or confidence in this area.

Is sensitive information consistently classified and governed across your organisation?

[ ] Controlled – Clear visibility with consistent governance and security controls. [ ] Partially Controlled – Some controls exist but are inconsistent, incomplete or difficult to enforce. [ ] Uncontrolled – Limited visibility, governance or confidence in this area.

Can you prevent confidential information being shared with unauthorised AI services?

[ ] Controlled – Clear visibility with consistent governance and security controls. [ ] Partially Controlled – Some controls exist but are inconsistent, incomplete or difficult to enforce. [ ] Uncontrolled – Limited visibility, governance or confidence in this area.

Can you monitor and audit how AI tools access and use sensitive business data?

[ ] Controlled – Clear visibility with consistent governance and security controls. [ ] Partially Controlled – Some controls exist but are inconsistent, incomplete or difficult to enforce. [ ] Uncontrolled – Limited visibility, governance or confidence in this area.

Get your AI risk profile

First Name

Last Name

Work Email

Organisation

Job Role

Phone

Assessment Complete

Thank you — here's what your score means

This initial assessment is based on the answers you provided and gives you an immediate indication of your security maturity across Identity, Network, Cloud and Data. It's intentionally designed as a high-level assessment.

Your automated score doesn't take into account:

Your organisation's size and complexity
Industry and regulatory requirements
Existing security architecture and technology investments
Cloud and application environment
Current and planned AI adoption
Risk appetite and business priorities
Any additional comments or context provided during the assessment

Treat your score as an initial indicator of maturity — not a comprehensive security assessment.

What happens next

Your assessment will now be reviewed by Principle Networks.

A Principle Networks security specialist will review your responses, including any additional context you provided.

Within 24–48 hours, we'll provide a more tailored interpretation of your results.

This will consider your organisation and environment in greater context and will help identify the areas we believe warrant the greatest attention.