About
Independent AI assurance, evidence first.
Alacrix is an independent AI risk, security, safety and governance assurance consultancy. We assess AI systems built by internal teams and external vendors, test the controls that matter and give regulated, high-trust and professional-service organisations an evidenced basis for deciding what to trust.
Purpose
Independent challenge before AI is trusted with real work.
Position
AI Risk. Security. Safety. Governance. Independently assessed, tested and evidenced.
Why Alacrix exists
Almost no one is independently checking the AI.
AI systems are typically assessed by the people who built them or by the vendor selling them. Policy advisers review documents, and conventional security testing was not designed for delegated authority, retrieved content and autonomous action.
Alacrix occupies the gap. We independently examine material risk, test the security, safety and governance controls, state the exposure that remains and report to the people who carry accountability for the outcome.
Law and accountancy remain priority sectors because confidentiality, professional judgement and auditability are fundamental. The same operating model applies wherever trust, resilience and accountable decision-making matter.
Positioning
AI that stands up to scrutiny.
Method
Understand the risk. Test the controls. Assess residual exposure. Decide.
Promise
An evidenced independent opinion, not reassurance.
Operating principles
How we work.
The commitments that shape every assessment we issue.
Start from observed practice
We examine the systems, permissions, workflows and decisions that exist, not the control environment people assume exists.
Connect consequence to control
Control strength follows the data, authority and potential impact of the use case rather than a generic checklist.
Stay outside the build
We do not build, resell or operate the systems we assess. Independence is the reason our conclusion carries weight.
Test the failure path
Expected operation is not enough. We examine misuse, error, escalation, containment and recovery.
Name residual risk
Exposure that remains is stated clearly, assigned to an owner and accepted through an accountable decision.
Leave the decision with you
We provide an evidenced assurance opinion. Approving, conditioning or declining deployment remains an accountable organisational decision.
Next step
Start with an independent view of exposure.
An independent assurance review establishes material risk, tested control strength, residual exposure and the evidence required to decide.