.png)
Lucent Search research | AI & Data Leadership Report 2025–2026
AI governance and leadership accountability
Does your organisation know who can approve, challenge, change or stop the use of AI?
AI governance concerns the decisions an organisation makes about how AI is developed, introduced and used. It includes the evidence required before deployment, the responsibilities that continue afterwards and the circumstances that require intervention.
Those decisions can involve business leaders, technology and data teams, risk, legal and operational functions. Effective oversight depends on how their responsibilities connect.
Lucent’s research found that AI governance, keeping up with legislative changes and ethical considerations were significant challenges for respondents. For executive teams, the findings raise questions about whether oversight is sufficiently clear, informed and practical.
RESEARCH FINDINGS
What the research found
88% found AI governance challenging
23.7% of respondents rated AI governance as very challenging, while 64.5% found it moderately challenging.
The finding identifies widespread difficulty within the surveyed group. It does not establish that respondents lacked governance arrangements or that those arrangements were ineffective.
It does provide a reason to examine how governance operates in practice: who makes decisions, what information they receive and how responsibilities are maintained.
82% found keeping up with legislative changes challenging
29% rated keeping up with legislative changes as very challenging, with a further 52.7% finding it moderately challenging.
This describes respondents’ experience of following legislative developments at the time of the survey. It is not a measure of legal compliance.
For leadership teams, the implication is to establish who monitors relevant developments, who assesses their significance and how necessary changes reach the people responsible for delivery.
Around 78% found ethical considerations challenging
23.7% rated ethical considerations in AI as very challenging, while 53.8% found them moderately challenging.
These responses highlight the judgement involved in developing and using AI. They do not identify specific incidents or establish whether organisations caused harm.
BUSINESS IMPLICATIONS
What this means for the business
Governance needs to support decisions throughout the use of an AI system.
An organisation may need to decide whether a proposed use is appropriate, what evidence is sufficient to proceed, which limitations users must understand and how performance will be reviewed. It also needs to establish what happens when circumstances change or concerns emerge.
A policy can describe expectations. Leadership teams need to understand how those expectations translate into approval, supervision and intervention.
The practical question is whether the people carrying responsibility have the information, expertise and authority to exercise it.
GOVERNANCE
Rebecca’s perspective: make oversight usable
I would want a leadership team to be able to explain how a consequential AI decision is made and revisited.
Who brings the proposal forward? Who challenges the assumptions? Who accepts responsibility for its use? What would cause the organisation to reconsider?
The answers should be understandable to the people developing the system and those using it. They should also reflect the significance of the decisions the system will support or make.
I would examine four areas.
Responsibility across development and use
Identify the business owner, the technical responsibilities and the people providing assurance.
Be clear about who authorises deployment, who accepts any remaining risk and who remains accountable once the system is in use. Where suppliers are involved, establish which responsibilities they hold and which remain within the organisation.
Changes in personnel or suppliers should not leave those responsibilities uncertain.
Evidence appropriate to the decision
Agree what decision-makers need to understand before they approve a particular use.
Depending on the application, that may include intended purpose, performance, limitations, data quality, security, potential effects on people and the arrangements for human judgement.
The depth of review should reflect the consequences of getting the decision wrong. Decision-makers also need to know where evidence is incomplete and what uncertainty remains.
Monitoring, challenge and intervention
Define how the organisation will identify problems after deployment.
Establish who reviews performance, how users raise concerns and who can require a change or suspend use. Human oversight needs sufficient expertise, time and authority to affect what happens.
Make the response to concerns explicit. People need to know that reporting an issue will lead to consideration and a decision.
Executive visibility and follow-through
Give the executive team a clear view of significant uses, unresolved concerns and decisions requiring its involvement.
Reporting should help leaders understand where responsibility sits, what has changed and whether agreed actions have been completed.
Where relevant obligations or internal standards change, assign responsibility for assessing the implications and implementing the response.
LEADERSHIP QUESTIONS
Questions for your leadership team
Use these questions when reviewing AI oversight or defining a leadership role.
Who owns each significant use of AI?
Is responsibility clear across development, deployment and ongoing operation?
Who has authority to approve or stop it?
Are those decision rights understood by the people involved?
What evidence supports approval?
Can decision-makers explain the limitations, uncertainties and consequences?
How can users challenge the system?
Is there a clear route for raising concerns and obtaining a response?
What happens when circumstances change?
Who assesses new information, supplier changes or relevant requirements?
What does the executive team need to see?
Does reporting make material concerns, decisions and overdue actions visible?
HIRING IMPLICATIONS
What this means for leadership appointments
An AI or data leader may need to establish governance arrangements, work within existing structures or improve how oversight connects to delivery.
The requirement should specify which responsibilities belong to the role and which are shared with business, technology, legal and risk leaders.
Assessment should examine how candidates have exercised judgement in relevant circumstances. What did they approve or challenge? How did they work with assurance functions? What happened when commercial pressure conflicted with unresolved concerns?
A useful discussion will establish their personal contribution, the evidence available and the consequences of the decision.
The organisation should also clarify the support the leader will inherit. Responsibility for governance needs to be matched by access to expertise, information and an effective route for escalation.
Lucent’s Leadership Requirement Diagnostic examines these relationships alongside the expected business outcomes. It helps define the judgement, experience and authority required for the appointment.
EXPLORE RESEARCH
Explore related findings
AI ownership, accountability and executive sponsorship
How responsibilities and decision rights connect across the organisation.
Scaling AI from pilot to production
The operational requirements involved in introducing and sustaining AI.
The skills AI and data teams need to deliver
The technical, interpersonal and organisational capabilities required across the team.
ABOUT
About this research
This page draws on the team structures section of Lucent Search’s AI & Data Leadership Survey Results 2025–2026, covering team size, geographic distribution and the discussion of team capabilities.
The findings describe the participating leaders’ circumstances at the time of the survey. They do not establish a preferred organisational structure or staffing ratio.
CONNECT
Clarifying leadership responsibility for AI?
You may be reviewing where oversight should sit, strengthening an existing function or defining the responsibilities of a new senior appointment.
Speak to Rebecca about the decisions the role needs to own, the expertise already available and the organisational relationships required for effective oversight.