Regulatory compliance and transparency
The published privacy policy does not describe the AI assistant, creating a blocker for screening decisions.
AIQURIS · The AI Lifecycle Management Platform
AI is moving faster than organisations can control it.
AIQURIS works out the risks, requirements and controls of every AI deployment, tracks whether they are in place, and keeps them current. So AI deployments go live faster, with a clear basis for every decision.
From one deployment to your entire AI portfolio. For SMEs and enterprises alike.
WorkspaceILLUSTRATIVEEvery deployment, with its risk, its status and what needs attention.
| Deployment | Owner | IMPACT+ | RISK+ | Status | Next action |
|---|---|---|---|---|---|
| Visual defect inspectionComputer vision · AI-024 | Manufacturing quality | Moderate | Moderate | Review due | Review changed contextDue 6 Oct · context change recorded |
| Candidate screening assistant ↗Decision support · AI-031 | Talent acquisition | Major | Critical | In assessment | Provide control evidenceDue 29 Sept · 24 required actions open |
| Credit origination affordability assistantDecision support · AI-009 | Risk and lending | Major | Critical | Reassessment | Provide testing evidenceDue 4 Oct · 3 controls open |
| Grid load forecastingPredictive AI · AI-052 | Network operations | Moderate | Moderate | In operation | Review on material changeCondition-based · last assessed 8 Aug |
| Contract summarisation copilotGenerative AI · AI-066 | Legal and compliance | Pending | Pending | Awaiting triage | Complete IMPACT+ triageRegistered 9 Sept |
| Internal knowledge chatbotGenerative AI · AI-018 | IT and operations | Minor | Not required | In operation | Scheduled reviewDue 15 Nov |
No sample deployments match your search.
Every deployment in one AI register, with its risk, its status and what needs attention. Control, deployment by deployment. Visibility across all of them.
The challenge
Deployments still stall in pilot, or go live with unmanaged risk.
Set requirements for AI in general. They don't say what a deployment needs, or whether it is in place.
A snapshot. Only one of many controls a deployment needs.
Covers selected aspects, after the fact. Shows what happens, not what should be controlled.
Manual and person-dependent. They don't scale, and go out of date when a deployment changes.
None of them works out what each deployment needs, or keeps it current.
It works out what your specific deployment needs, maps it against what you already do, and shows what's missing, at every stage of the AI lifecycle.
Because it starts from the risks of the actual deployment, it is:
Why it is hard
AI risk is hard to tell. It cuts across disciplines, differs by deployment and changes over time, and the controls differ with it. Below, one system in four sectors: same technology, four different risk profiles.
| Risk domain | Financial servicesAdvisers answering client product questions | HealthcareClinicians checking treatment guidelines | Defence & critical infrastructureControl-room operators querying operating procedures | ManufacturingTechnicians querying machine maintenance manuals |
|---|---|---|---|---|
| Safety | Negligible | Critical | High | Critical |
| Security | High | High | Critical | Moderate |
| Performance | Moderate | High | High | High |
| Legal compliance | Critical | High | High | Moderate |
| Ethics | Moderate | High | Moderate | Low |
| Sustainability | Low | Low | Low | Low |
| Deepest controls on | Legal compliance | Safety | Security | Safety |
How it works
A guided intake captures everything that matters about your deployment. The engine then applies the standards, regulations and policies AIQURIS maintains, and works out the risks, requirements and controls it needs to be safe, perform and comply.
The missing link
Governance sets the rules. Teams build and run the deployments. Legal, cybersecurity, risk and the business each see one part. AIQURIS sits in between and connects them all: which rules apply to each deployment, where it stands, and who acts next.
How the engine worksExample
One real deployment, traced from risk to evidence. AIQURIS assessed a candidate screening assistant before its recruitment pilot went live.
Protected attributes or proxy signals could unfairly exclude candidates from progressing.
Protected attributes must not influence results.
Human review before any rejection. Testing that protected and proxy attributes do not affect scores.
Test results against defined thresholds. Review records showing that people made the decisions.
The published privacy policy does not describe the AI assistant, creating a blocker for screening decisions.
Interface defaults, hidden filtering or cut-offs could turn decision support into automated exclusion.
Ambiguous criteria and extraction errors could produce unreliable scores and ranking errors.
Applicant data in prompts, responses and logs requires verified minimisation, retention and vendor controls.
RECOMMENDATIONProvide evidence for the critical controls before go-live.
DECISIONTalent acquisition held the pilot until the evidence is in.
Summarised from an AIQURIS assessment of a real deployment. Ratings are before mitigation (inherent risk).
Lifecycle
One connected record across the deployment's life, shared by every function involved. When the deployment or its governing sources change, requirements, controls and evidence are updated together.
EXAMPLE The approved job criteria changed after go-live. Requirements reassessed, controls reopened, evidence updated, decision on continued operation updated.
Every product includes reassessments. CONTROL+ adds defined change triggers.
Products
One platform, three levels of depth. Start with IMPACT+ and go deeper only where the risk calls for it, so effort and cost match the deployment.
Who could be harmed, and how badly.
YOU RECEIVEHow the harm could occur, and which controls address it.
EVERYTHING IN IMPACT+, PLUSAI Assurance: the controls in detail, the remaining risk, the evidence.
EVERYTHING IN RISK+, PLUSNo AI governance team needed. Start with one deployment and pay per deployment.
One AI register for the whole portfolio. CONTROL+ where the risk calls for it, with expert-supported setup.
Why AIQURIS
AIQURIS was founded by Dr Andreas Hauser and Dr Martin Saerbeck. Engineering, certification, digital systems and mathematical modelling, now applied to AI. Seven years developing the AIQURIS methodology, the early years within TÜV SÜD’s global AI quality practice. AIQURIS is independent.
Selected standards contributed to. Meet the founders and explore their standards work ↗
“The solution AIQURIS presented is the first I know of in the market that closes the gap by combining AI risk management and requirements from technical standards in a tool-based way.”
IMDA Spark Company · AI Verify Foundation member
The team sits on the committees that write the AI standards.
Get started
The AI Deployment Assessment Sprint is guided by AIQURIS from start to finish. We prepare the assessment; your team gives five hours across three sessions. Within three weeks, you have one real deployment assessed at RISK+ depth, and actions with named owners.
Choose a deployment, in planning or already in production. AIQURIS then screens its impact and prepares the risk profile, requirements and controls.
The accountable stakeholders in the room. Validate the risks, challenge the controls, agree what closes the gaps.
Leave with an assessment record and an action list with named owners. Run the next deployments yourself on the platform.
Questions first? Contact info@aiquris.com
Dashboard and lifecycle views are illustrative.