Industries
Governance is sector-specific. So is our method's application.
The controls that satisfy an airworthiness auditor are not the controls that satisfy a model risk validator or a privacy commissioner. PMAIS is configured per sector against the obligations that actually apply.
Sectors
Sectors we have delivered in.
Nine sectors where we have direct delivery experience — and the same method applies wherever AI decisions have to be explained to someone else.
Aerospace
Safety-critical systems certification context
AI introduced near safety-critical systems inherits the certification burden of those systems. We structure deployments so change control, traceability, and design-assurance expectations are satisfied by evidence generated during delivery — not reconstructed afterward for an airworthiness or supplier audit.
View sector →Banking & capital markets
Model risk management and supervisory explainability
Supervisors already have a mature model-risk vocabulary. We express AI systems in that vocabulary — tiering, independent validation, effective challenge, decision-level explanation — so credit, pricing and surveillance deployments clear internal validation and supervisory review without a parallel governance stack.
View sector →Health
Clinical safety, patient privacy and device classification
Anything that informs care carries a clinical risk-management duty and a privacy duty at the same time. We determine device classification early, build the hazard log and safety case alongside delivery, and specify clinician oversight at every point where a model output can reach a patient pathway.
View sector →Telecommunications
Network reliability, subscriber privacy and lawful access
Operators automate at a scale where a bad model decision becomes an outage. We bound blast radius, stage rollout and rehearse rollback for every automated network action, while keeping subscriber data inside the privacy and lawful-access constraints of each jurisdiction served.
View sector →Insurance
Fairness testing, actuarial validation and conduct supervision
Underwriting and claims decisions are the most closely watched AI applications in insurance, because the harm from an unfair model is individual, documented and litigable. We evidence fairness testing and actuarial review rather than asserting them, and keep a reviewable trail behind every customer-affecting decision.
View sector →Transportation & logistics
Operational safety cases and transport regulator scrutiny
Scheduling, dispatch and asset-condition models sit close to physical safety and to duty-of-care rules on hours, load and maintenance. We build the operational safety case with the deployment, so a transport regulator or insurer can see who approved an automated decision, on what basis, and how it can be overridden.
View sector →Manufacturing & industrial
Product quality, worker safety and OT boundary control
On the plant floor the risk is a model that quietly changes a quality disposition or reaches into operational technology. We keep the IT/OT boundary explicit, tie model-influenced quality decisions to the existing quality management system, and treat worker-safety impacts as a delivery gate.
View sector →Research & academia
Research integrity, data agreements and reproducibility
Institutions face breadth rather than depth: many small deployments, many data agreements, and a reputational cost to any integrity failure. We give review boards one consistent way to assess AI components, and make AI-assisted findings disclosable and reproducible by default.
View sector →Government & public sector
Algorithmic impact assessment and public accountability
Public deployments answer to the citizen as well as the auditor. Impact assessment, model provenance and procurement records are delivery artifacts, not communications afterthoughts — and every automated decision needs a staffed route to challenge it.
View sector →Beyond these sectors
Not in a regulated sector? The method still applies.
Phases, gates and human-in-the-loop design are not a compliance ritual — they are how any organization deploys AI it can explain, defend and keep running. Retail, energy, professional services, education, non-profit and technology teams use the same spine: named approvers, documented controls, retained evidence. If your AI decisions affect customers, money, safety or reputation, PMAIS fits.
Jurisdictions
Worldwide coverage, one governance spine.
Multinational clients need a single control framework that satisfies the strictest applicable regime while remaining operable everywhere else.
Canada
AIDA readiness, PIPEDA and provincial privacy regimes, federal directive on automated decision-making.
United States
NIST AI RMF, sector supervision including SR 11-7, and state privacy and automated-decision rules.
European Union
EU AI Act classification and conformity, GDPR interaction, and post-market monitoring duties.
United Arab Emirates
UAE AI strategy and charter expectations, ADGM and DIFC data protection regimes, and sector regulator guidance.
Australia
Privacy Act reform, APRA prudential standards and the voluntary AI safety standard.
Rest of world
UK, Singapore, Japan, Gulf and Latin American regimes mapped onto the same control framework wherever you operate.