AI Assurance platform
Ituti AI Assurance is our platform for assessing AI systems against the EU AI Act, UK requirements, ISO/IEC 42001 and the NIST AI RMF. Your team and ours work in one workspace, and every finding is backed by evidence an auditor can follow.
Example system
Personal loan decisioning
Retail Credit · in production
Inherent risk
66 / 100
Basis
Art. 6(2), Annex III(5)(b)
Control readiness
What it does
Each AI system gets one record: what it is, how it is classified, which obligations apply, how it measures up against each control, and what happened after it went live. Regulators, auditors and your own risk committee all see the same evidence.
Screens the eight prohibited practices, the Annex III high-risk uses and the transparency triggers, and returns a risk tier with the article-level reasons behind it.
EU duties for providers and deployers, plus the UK picture: the AI principles, UK GDPR, the Equality Act, and FCA Consumer Duty, SM&CR and PRA SS1/23 where they apply.
Up to 39 controls, filtered to the ones that apply to the system, each cross-referenced to EU AI Act articles, ISO/IEC 42001 clauses and NIST AI RMF functions.
Every control carries its evidence, with readiness by framework and a prioritised list of gaps. An optional review by Claude flags evidence that looks weak or missing.
A printable assurance report and model card for each system, ready for a risk committee, an internal audit or a customer’s due-diligence questionnaire.
Track performance and fairness metrics against agreed thresholds, and log incidents, flagging any that may need reporting to the authority under Article 73.
How an assessment runs
Each step builds on the last, so nobody answers the same question twice.
What the system does, where it is used, who it affects and how automated it is. This sets the inherent risk score.
EU AI Act tier with its reasons, and the obligations that follow in the EU and the UK.
Each applicable control marked met, partial or not met, with the evidence recorded against it.
Coverage by framework, the gaps to close, and a model card you can share.
Metrics and incidents logged against the same record, so the evidence stays current.
Built for regulated teams
Each client organisation has its own workspace. Your team can be as involved as you like: we can run the assessment for you, or your people can complete it and we review.
Ituti assessor
Leads the assessment, reviews evidence and signs off the report for the clients assigned to them.
Client admin
Your AI or risk lead. Adds systems, invites colleagues and decides who can edit what.
Contributor
A system owner or data scientist. Edits only the systems assigned to them and records incidents.
Viewer
Internal audit, compliance or the board. Reads every system, report and model card.
Where it fits
Ituti AI Assurance gives you the regulatory classification, control assessment and evidence pack. Where you need automated model monitoring and enterprise risk management at scale, we implement IBM watsonx.governance and OpenPages, and the assessment carries straight across.
Questions
Bring us the AI system that worries you most. We will classify it, assess it with your team and hand you the report.