Is your organisation governing AI — or just using it?
Artificial intelligence is already shaping who gets hired, how performance is assessed, and who gets dismissed across UK workplaces. In most organisations, that influence is outpacing the governance designed to oversee it.
This white paper sets out why that gap matters — legally, financially and culturally — and what structured oversight actually looks like in practice.
Download free
AI-Influenced Decision Risk
Enter your email address and we'll send the PDF directly to your inbox. No registration wall.
After downloading, you'll receive two short follow-up emails over the next fortnight — one practical scenario, one question worth asking your board. Nothing more unless you get in touch.
What's inside
35 pages. Everything your board and HR team needs.
Written for leaders who need to understand AI-Influenced Decision Risk as a governance matter — not a technology matter.
The central argument
The primary risk of AI in workforce systems is not adoption — it is ungoverned algorithmic influence over consequential employment decisions. This paper reframes AI readiness as a governance question, not a technology question.
The legal landscape — fully updated for 2025/26
Equality Act 2010, Data (Use and Access) Act 2025, Employment Rights Act 2025, and the EU AI Act obligations live from August 2026. What each one means for HR directors and boards in plain language.
The financial exposure — in real numbers
Employment tribunal awards without a cap. ICO fines up to £17.5 million. What the removal of the unfair dismissal compensation cap from January 2027 means for every AI-influenced decision your organisation makes from now.
Agentic AI — the governance frontier most boards don't know exists
What happens when AI doesn't just recommend, but acts. A concrete scenario: 2,400 applications processed, 1,980 rejected automatically, 12 disability discrimination claims. No audit trail. ICO investigation opened.
Why a human in the loop isn't enough
The research on automation bias, diffusion of responsibility and ceremonial review that explains how organisations with good intentions still end up exposed. What meaningful human review actually requires.
The inclusion dimension
What algorithmic systems are trained to reward, what they systematically penalise, and why neurodivergent candidates and employees are disproportionately affected. The legal and strategic case for embedding adjustment capability into AI-supported processes.
The AI-Influenced Decision Maturity Model
Four levels of governance development, from informal adoption to embedded culture. Where most organisations actually are versus where they believe they are — and what the gap costs them.
A five-pillar leadership framework
Visibility, accountability, decision integrity, inclusion capability, and continuous oversight. Practical and specific — not generic. A framework boards can actually apply.
Included in the paper
The agentic AI scenario that changes the board conversation
An agentic HR system configured to reduce time-to-hire processes 2,400 applications over six weeks. It rejects 1,980 candidates automatically — before any human reviews a single decision. The scoring model, trained on historical data, assigns lower scores to employment gaps and non-linear career histories. Three months later: twelve disability discrimination claims. An ICO investigation. No audit trail. No documented human review. This is not a hypothetical. It is a realistic combination of deployments already operating across UK employers.
Self-assessment diagnostic
Where does your organisation actually sit?
The diagnostic is included in the white paper. Two versions — one for boards, one for HR directors. Scored, with guidance on what your results mean and what to do next.
For boards & NEDs
Can you name the executive responsible for governance of AI-influenced workforce decisions?
Is AI-Influenced Decision Risk formally recorded on your risk register?
Can you confirm that human review of AI decisions is substantive — not a procedural sign-off?
Do you know whether the organisation uses agentic AI systems taking autonomous actions in HR?
Are you aware of how ERA 2025 and DUAA 2025 affect your legal exposure?
For HR Directors & CPOs
Have you mapped all AI systems influencing recruitment, performance, absence or disciplinary decisions?
Are adjustment pathways for neurodivergent employees embedded in AI-supported workflows — not external exceptions?
Do you monitor outcome data for patterns of disproportionate impact by protected characteristic?
Can you produce documented decision rationale if required in a tribunal investigation?
Does your vendor contract include audit access and transparency obligations?
Most organisations, when they work through these questions honestly, find themselves at Level 1 or 2 of the Maturity Model — regardless of what they believe about their governance. The full diagnostic with scoring is in the white paper.
The maturity model
Four levels of AI governance development
Governance maturity is not defined by technological sophistication. It is defined by alignment between algorithmic influence and organisational oversight.
AI systems operate without board visibility or structured oversight. Vendor assurances accepted without employer-level interrogation. Most organisations are here.
Influence mapped and acknowledged. Policies drafted. Governance inconsistent. Risk of performative compliance — appearance without architecture.
Named accountability. Formal risk register integration. Impact assessments embedded. Documentation standards defined. Compliance capability, not just awareness.
Oversight is routine. Adjustment integrated. Override patterns monitored. Continuous review normalised. Governance is institutional identity.
Who this is for
Written for leaders, not technologists
Effective AI governance does not require technical expertise. It requires the right questions, the right evidence, and governance structures that function in practice.
Boards & Non-Executive Directors
For those who want to understand AI-Influenced Decision Risk as a governance and stewardship matter — and ensure oversight mechanisms are proportionate to algorithmic reach.
HR Directors & Chief People Officers
For those navigating the intersection of AI tools, employment law and inclusion obligations — and who need to be able to evidence their governance position under scrutiny.
General Counsels & Legal Teams
For those advising on ERA 2025, DUAA 2025 and EU AI Act compliance in employment contexts — and the cumulative exposure when legal channels operate concurrently.
Any organisation where AI shapes workforce decisions
Off-the-shelf recruitment platforms, performance management software and workforce analytics tools all carry these obligations. You don't need to build bespoke AI to be in scope.
Download the white paper — free
35 pages. Self-assessment diagnostic for boards and HR directors. Updated for the latest legislative and regulatory developments. Sent directly to your inbox.
Free download
AI-Influenced Decision Risk
Enter your email and we'll send the PDF immediately.
No registration wall. Two follow-up emails. Unsubscribe any time.
About the author
Lucy Smith
Founder & Director, Inclusive Change Ltd and Inclusive Change at Work CIC
Lucy Smith is a change management specialist, social entrepreneur and inclusion consultant with over two decades of experience leading organisational transformation across public, educational and social sector organisations.
Her career includes senior change leadership roles within the National Crime Agency and the University of Bristol, where she led large-scale transformation programmes, developed organisational change capability, and advised senior leaders on the people impacts of strategic change.
Drawing on expertise in governance, workforce development, neurodiversity and organisational culture, Lucy's work focuses on helping organisations balance innovation with ethical practice — ensuring that technological advances strengthen rather than undermine inclusion, equity and human potential.
How Inclusive Change can help
- AI governance maturity assessments
- Board briefings on AI-Influenced Decision Risk
- Decidr crisis simulations for boards and executive teams
- Leadership development on influence literacy and AI governance
- Equality impact assessment for AI-supported processes
- Implementation support for governance architecture
Inclusive Change Ltd
The Brightwell, Bradbury House
Wheatfield Drive
Bradley Stoke, Bristol
BS329DB
Reg no: 12412464
Copyright 2026 Inclusive Change Ltd
VAT NO: 352 1564 17
ICO Reg: ZB081779
UK Register of Learning Providers: 10090652