Your organisation is deploying AI.
But can your people actually govern it?
The EU AI Act requires demonstrable human oversight of AI systems. Most organisations have built the frameworks and policies. Few have measured whether their people are psychologically equipped to use them when it matters.
If your organisation deploys AI systems that affect people — in financial services, healthcare, or the public sector — and you need to demonstrate that the people responsible for oversight can actually exercise it, this is for you.
Compliance programmes build processes.
They don't build the people who run them.
Organisations are investing heavily in AI governance frameworks, policies, and training — yet remain fundamentally blind to whether their leaders and teams feel safe to challenge AI outputs, escalate concerns, or override systems when it matters.
Traditional maturity assessments tell you where you stand on a generic five-stage model. They don't tell you whether your people can exercise the judgement that the EU AI Act explicitly demands.
The gap is not awareness. The gap is measurable readiness. And readiness — unlike awareness — can be assessed, quantified, and acted upon.
The problem is not that psychological and cultural factors are unmeasurable. The problem is that they have not yet been translated into decision-relevant, governance-ready indicators. That is precisely what we do.
Why the gap persists
Automation bias is invisible
People trust AI outputs more than their own judgement — not because they are told to, but because it is cognitively easier. This is a measurable psychological phenomenon, not a training gap.
Psychological safety is assumed
Most organisations assume their people feel safe to challenge AI. Very few have measured it. In practice, the conditions that enable genuine challenge are rare — and fragile.
Governance stops at documentation
Policies describe what should happen. They do not predict what will happen when a human operator faces a high-confidence AI recommendation under time pressure.
Readiness has never been measured
Until now, there has been no structured, practitioner-ready way to assess whether the people in oversight roles are genuinely equipped to exercise the judgement the EU AI Act requires.
The compliance window is narrowing
The EU AI Act's obligations are arriving in phases. For organisations deploying high-risk AI systems, the most consequential deadline is August 2026 — and the work required to meet it cannot be compressed into the final months.
Article 4 is enforceable
Organisations must ensure a sufficient level of AI literacy for all staff dealing with AI systems. Generic e-learning will not meet this standard.
Certain AI practices become prohibited
Organisations must have identified and discontinued any systems that fall within the Act's prohibition categories.
Complete compliance required — date under revision
Original deadline: August 2026. The Digital Omnibus proposal is currently in EU trilogue — dates remain subject to change.
MEPs propose 2 December 2027 for high-risk AI systems listed in the regulation (biometrics, critical infrastructure, education, employment, essential services, law enforcement, justice and border management).
⚠ Not yet confirmed. Organisations should continue preparing against August 2026.
From diagnosis to embedded capability
We follow a structured methodology that begins with measurement and ends with demonstrable governance capability. Every step is grounded in evidence — not assumption.
Discovery Conversation
A structured conversation to understand your current governance posture, key risks, and whether our diagnostic approach is the right fit.
Governance Diagnostic
A structured, research-informed assessment of whether the people in your organisation are genuinely equipped to exercise AI oversight.
Advisory & Architecture
Based on what the diagnostic finds, we advise on the specific interventions required — deployer readiness, literacy architecture, or governance redesign.
Implementation & Review
We support implementation, measure progress through re-assessment, and ensure governance capability outlasts any single programme.
Four layers of AI governance. Only one is about people.
AI governance conversations tend to stop at policy and process. The question most organisations haven't yet asked is whether the people responsible for oversight can actually exercise it — in real conditions, under real pressure, against the pull of systems that are designed to be trusted.
What the system does
The technical infrastructure surrounding AI — how models are monitored, how data is governed, how outputs are tested and audited before they reach decision-makers.
What the organisation requires
The documented obligations — policies, registers, reporting lines, and regulatory filings. The paper record of how governance is supposed to work.
Who is accountable
The authority structures that determine who owns which decisions, who can escalate, and who is ultimately responsible when something goes wrong. Roles on paper.
Whether people can actually act
The human dimension — whether the individuals in those roles are genuinely equipped to question, override, escalate, and take responsibility. This is the layer most organisations have never assessed. It is also the layer regulators are beginning to require.
Where we work — Art. 4, 14, 26One entry point. One honest question.
Can your people actually govern AI — and can you prove it? The Diagnostic answers that question. What we do next depends entirely on what it finds. We do not arrive with a pre-set agenda. We arrive with a structured way to find out the truth.
Not a training provider
We do not deliver e-learning or awareness programmes. We specify the architecture others build — and measure whether it is producing genuine capability.
Not a maturity model
We don't place you at "Level 2" on a generic scale. We identify specific gaps, in specific roles, against specific regulatory obligations — then tell you what to do about them.
Not a compliance checkbox
Documentation tells a regulator what you intended. Our work tells you — and them — whether your organisation can actually deliver it. That is the distinction the EU AI Act enforces.
Governance Diagnostic
A structured assessment of whether the people in your organisation are genuinely equipped to exercise AI oversight — not in theory, but in the conditions they actually face. We examine the attitudes and behaviours that research consistently links to governance effectiveness: the willingness to question, the readiness to escalate, the sense of personal accountability for decisions made with AI.
The output is not a maturity score. It is a clear picture of where human oversight is likely to hold — and where it is likely to fail — mapped against what the EU AI Act requires your organisation to demonstrate.
Book a Discovery Conversation →- A role-specific readiness profile across your oversight-critical population
- Clear identification of where oversight is likely to hold and where it is likely to fail
- A prioritised advisory agenda — the specific interventions the evidence recommends
- Regulatory-ready documentation aligned to Articles 4, 14, and 26
Advisory services — determined by what the diagnostic finds
EU AI Act Deployer Readiness
Most EU AI Act attention focuses on providers — the organisations building AI systems. Article 26 places equally significant obligations on deployers — the organisations using them. Most are not ready.
When the diagnostic reveals gaps in human oversight assignment, vendor governance, or incident response capability — we advise on the deployer compliance programme that closes them: obligation mapping, oversight role specification for each high-risk system, vendor due diligence framework, and an incident response protocol aligned to the Act's 72-hour notification requirement.
AI Literacy Architecture
Article 4 has been enforceable since February 2025. It requires a sufficient level of AI literacy — a standard that generic e-learning cannot meet, and that most organisations cannot yet demonstrate.
When the diagnostic reveals that literacy gaps are undermining the capacity to challenge, escalate, or override — we specify the role-differentiated architecture your L&D team builds against: needs analysis by role, tiered programme specification, and an effectiveness measurement framework. We specify it. You build it.
Retained Advisory
AI governance is not a one-time project. Regulation develops. AI systems evolve. People move. The human factors that determine oversight effectiveness change as organisations restructure.
Most advisory relationships end when the report is delivered. Ours are designed to continue — with periodic re-assessment, governance health checks, emerging regulatory interpretation, and on-call support when something happens that can't wait.
Rigorous. Independent. Evidence-led.
The Responsible AI Center is an independent advisory firm. We are not reselling a platform, certifying a standard, or advocating a vendor. We are here to give your organisation an honest picture of its human oversight capacity — and a practical path to improve it.
Measurement before recommendation
We diagnose before we prescribe. No engagement begins with a pre-packaged solution. Our governance diagnostic produces the evidence that every subsequent recommendation is built on.
Regulatory precision
We work to Articles 4, 14, and 26 of the EU AI Act and ISO/IEC 42001:2023 Clauses 6.2, 7.3, and 8.4. Our outputs are designed to support audit committee reporting and regulatory scrutiny.
Sector-specific expertise
We work across financial services, healthcare, public sector, and professional services — wherever high-risk AI is deployed and human oversight is a regulatory obligation.
Intervention that sticks
We embed governance capability to outlast any single programme. Re-assessment cycles, embedded accountability, and governance mechanisms that survive leadership changes.
Built for the deadline
Our methodology is calibrated for organisations that need to demonstrate human oversight capacity within a constrained window — without sacrificing rigour.
Independent.
Evidence-led. Built for the practitioner.
The Responsible AI Center is an AI governance diagnostics and advisory firm based in Brussels. We work with organisations subject to the EU AI Act who need to demonstrate not just technical compliance, but that their people are psychologically equipped to exercise genuine human oversight of AI systems.
Our focus is the gap between what organisations have built — frameworks, policies, technical controls — and whether the people operating those structures are actually ready to use them. That gap is assessable. It is quantifiable. And it is the gap regulators will scrutinise.
People govern AI. We make sure they can.
Mulya works at the intersection of enterprise technology, risk advisory, and AI governance. His practice focuses on helping boards, executives, and control functions turn regulatory obligation into operational capability — embedding responsible AI into the way organisations design, deploy, and oversee intelligent systems.
He brings almost two decades of experience from KPMG, IBM/Kyndryl and Microsoft, combined with deep specialisation in the EU AI Act, ISO/IEC 42001:2023, and human oversight frameworks. His work spans highly regulated sectors like financial services, healthcare and the public sector.
Mulya collaborates with academic partners on AI governance research. Based in Brussels, he advises organisations across the EU.
Research-informed. Developed with academic partners.
Our advisory approach is informed by ongoing research into the psychological and organisational conditions that determine whether AI governance works in practice. We collaborate with academic partners to develop the constructs that inform our diagnostic tools.
Psychological safety as a governance condition
Automation bias susceptibility in oversight roles
Construct measurement for AI oversight readiness
Translation of regulatory obligation into observable behavioural indicators
This research is not academic for its own sake. It exists to ensure that every recommendation we make is informed by structured inquiry — not assumption, not convention, and not what worked in a different regulatory context.
Thinking on AI governance and human oversight
Research-grounded perspectives on the EU AI Act, governance design, and the psychology of responsible AI adoption. Written for practitioners, not academics.
Independent by design. Collaborative by nature.
The Responsible AI Center works with a small number of academic, institutional, and practitioner partners whose work is directly relevant to the governance challenges we address.
Research in AI Governance & Psychological Readiness
Our governance diagnostics are informed by joint research into the psychological and organisational conditions required for responsible AI oversight. This work bridges academic inquiry and boardroom practice — grounding our approach in ongoing research while keeping it actionable for practitioners.
Psychological safety as a governance condition
Construct measurement for AI oversight readiness
Translation of regulatory obligation into observable behavioural indicators
Governance, Risk & Compliance Specialists
We maintain working relationships with a select group of GRC practitioners, legal advisors, and HR transformation specialists across the Netherlands, Belgium, and the wider EU. These relationships allow us to refer clients to complementary expertise — and to bring relevant perspectives into our own engagements where appropriate.
If you work in AI law, organisational psychology, or regulatory compliance and believe there is a meaningful basis for collaboration, we welcome the conversation.
Reach out to explore →Substantive, not decorative
Every collaboration we enter into has a clear intellectual or practical purpose. We do not maintain partner lists for appearance — we work with people whose expertise makes our work sharper.
Independence preserved
Collaboration does not compromise our independence. We do not take referral fees, endorse products, or allow commercial relationships to influence our governance diagnostics or advisory positions.
Client interest first
When we refer clients to partner specialists, it is because those specialists are genuinely the right resource — not because of any commercial arrangement. Our clients' interests are the only criterion.
Start with a conversation.
Every engagement begins with a Discovery Conversation. Just an honest exchange about where your organisation stands, what the EU AI Act requires, and whether we are the right fit to help close the gap.
Message Received
Thank you for reaching out. We will be in touch within two working days.
