.png)
What Does the Board Actually Need to Decide About AI, Data and Technology?
Boards do not need to run technology programmes. They do need to make decisions that determine whether those programmes have a realistic chance of succeeding.
That distinction becomes more important as AI, data and technology move further into business strategy, operating models, investment and risk.
The board may never decide which AI platform to buy or how the data architecture should be designed. It should understand what the organisation is trying to achieve, who is accountable for delivering it, whether the leadership team has the capability and authority required, and where the board itself needs to make a judgement.
The conversation reinforced something I see repeatedly in executive search and leadership work: boards are being asked to make increasingly consequential technology decisions without necessarily being clear about which decisions genuinely belong in the boardroom.
Directors do not need to become technologists. They do need greater clarity about the decisions only the board can make.
Decide what the organisation expects AI and technology to change
Technology strategies can become disconnected from the business surprisingly easily.
A board sees a significant investment request for AI, data infrastructure, enterprise resource planning, cloud or digital transformation. The programme has objectives, milestones and a business case.
But there is a more fundamental question:
What should be materially different in the organisation if this investment succeeds?
The answer needs to be expressed in business terms.
Perhaps the organisation needs to improve operating reliability, reduce the cost of serving customers, make better investment decisions, create a new product, increase productivity or manage risk differently.
Once that outcome is clear, the board has a much stronger basis for challenging the technology strategy.
Without it, discussion can drift towards activity: pilots launched, platforms deployed, people hired and programmes delivered. Those things may all be necessary, but they are not the business outcome.
Know who is accountable
AI and data make accountability particularly difficult because responsibility often crosses several executive roles.
The Chief Information Officer (CIO) may own enterprise technology. The Chief Technology Officer (CTO) may own engineering and architecture. The Chief Data Officer (CDO) or Chief Data and Analytics Officer (CDAO) may own data, analytics and parts of the AI capability. A Chief AI Officer (CAIO) may own an enterprise AI agenda.
Individual business executives still own the commercial or operational results.
The board needs enough clarity to understand where accountability actually sits.
If AI investment is failing to deliver value, who is expected to resolve it? If poor data is preventing progress, which executive owns the problem? If a business function will not adopt a new way of working, who has the authority to intervene?
And where responsibility is distributed, who resolves disagreements between functions?
These questions become more important as AI moves out of innovation teams and into core operations.
Lucent Search's research found that 87% of AI and data leaders experience difficulty with cross-departmental collaboration. Many AI outcomes now depend on technology, data and business functions working together.
The board does not need to design every reporting line. It should be able to see whether the accountability model makes sense.
Make sure authority follows accountability
This is easily missed when organisations appoint senior technology and AI leaders.
An executive is given an ambitious role and significant accountability, but many of the decisions required to deliver it remain elsewhere.
The AI leader needs technology investment controlled by the CIO. The CDO needs business functions to change how they manage data. The CIO is expected to transform the organisation while individual business units control priorities and budgets.
The person may have the right title and still have very little ability to change the system around them.
Before approving a senior appointment or major transformation, the board needs to understand which decisions the executive can make, which they can materially influence and which depend on the cooperation of peers.
Lucent's research found that 71% of AI leaders struggle to secure sufficient budget, while 88% find AI governance challenging.
Alongside high levels of cross-functional friction, those findings point to a problem boards should recognise: executives can be held accountable for results without controlling enough of the investment, decisions or organisational support required to deliver them.
Decide where the organisation will tolerate risk
AI, data and technology investment inevitably involves uncertainty.
The board's role is to understand which risks the organisation is prepared to take, which require stronger controls and which cannot be accepted.
For AI, that may include customer outcomes, operational reliability, privacy, model behaviour, cyber risk, regulation and third-party providers.
For major technology transformation, the risks may include resilience, programme failure, business disruption or dependency on legacy systems.
The useful board discussion therefore extends beyond compliance.
Which AI uses would materially change the organisation's risk exposure? Which systems cannot tolerate failure? Which decisions require human oversight? Where does the organisation need stronger assurance before scaling?
Governance should make those choices clear rather than simply create additional approval stages.
This is particularly relevant in regulated, asset-intensive and heavily governed organisations, where technology decisions can affect operational resilience as well as commercial performance.
Look beyond the first AI pilot
Boards often become involved when a significant technology investment is first approved.
AI creates a slightly different problem because experimentation can be relatively easy to fund. Scaling is much harder.
A successful pilot may require further data work, systems integration, infrastructure, governance, people and changes to existing business processes before it can operate reliably across the organisation.
That can make the second investment decision more important than the first.
The board needs to understand how the organisation decides which experiments deserve further capital and which should stop.
Who makes that decision? What evidence is required? Can funding move quickly enough when something works? Is the board seeing a collection of experiments, or a coherent investment strategy?
If an organisation can fund pilots but cannot move successful work into normal operations, the constraint may sit in the way investment and operational decisions are made rather than in its ability to innovate.
Understand whether the operating model can deliver the strategy
A credible strategy and a strong executive are not enough if the organisation underneath them cannot deliver.
The board should understand the fundamentals.
Where does specialist AI capability sit? Who owns the data? How do technology teams and business units work together? Who prioritises competing use cases? How does promising work move into production? Which executive owns adoption?
There is no universal operating model.
Some organisations need a strong central AI capability. Others will embed people closer to individual businesses. Many will use a hybrid.
The board does not need to design that structure in detail. It does need enough visibility to know that the organisation has made deliberate choices about how the work will get done.
Without that clarity, progress often depends on personal relationships, informal sponsorship and individual executives negotiating their way through the organisation. That becomes increasingly fragile as investment grows.
Treat leadership capability as a board decision
This is where board oversight and executive appointments meet.
The board needs to understand whether the current leadership team can deliver the strategy it has approved.
That requires more than asking whether the CIO is performing well, whether the CDO is credible or whether the organisation should create a new AI title.
Look at the job the organisation now needs someone to do.
Does the CIO have the enterprise influence required for the next stage of technology transformation?
Can the CDO or CDAO move from improving the data estate into creating measurable business value from it?
Does AI require a dedicated executive, or can the existing CIO, CTO or CDAO lead it effectively?
Is there enough leadership beneath those executives to prevent already broad roles becoming unmanageable?
These are capability and organisational design decisions. They are best made before a search begins, rather than after candidates have entered the process.
Lucent's research also found that nearly 63% of AI and data leaders planned to move roles within the following 12 months. That statistic cannot tell us why any individual intends to leave, but it makes leadership continuity a legitimate board concern when AI and data have become central to the strategy.
Boards should know which executives have become critical to execution, what would happen if they left and whether the organisation has created roles in which strong people can realistically succeed.
Digital fluency means knowing how to challenge
Boards do need greater fluency in AI, data and technology. They do not all need deep technical expertise.
A digitally fluent board understands enough to ask sensible questions and recognise when the answers are weak.
Directors should be able to distinguish between a technology output and a business outcome. They should understand the difference between an AI pilot and something operating reliably in production. They should know that poor data can limit AI regardless of how sophisticated the model is.
They should also recognise when an apparently technical problem is actually one of unclear accountability, inadequate investment or organisational design.
That level of understanding materially improves the quality of board challenge.
Where specialist technical depth is required, the board can bring it in. Judgement should not simply be handed to whichever person in the room understands the technology best.
Board composition should reflect the decisions the organisation now faces
For some organisations, appointing a non-executive director (NED) with recent technology, data or transformation experience will materially strengthen the board.
For others, the existing board may have sufficient capability if directors are prepared to develop their understanding and can draw on credible external expertise when required.
Start with the decisions the board now faces.
If AI and technology have become central to operating performance, capital allocation and risk, can the current board challenge those decisions confidently?
If not, there is a capability gap.
That does not automatically require another permanent board appointment. It does require a deliberate response.
The board's role is decision quality
Boards do not have to choose between governing transformation and leading it.
They provide oversight, but they also make a relatively small number of decisions that shape whether AI, data and technology can create value.
They decide which outcomes are important enough to invest in. They make sure accountability is clear and test whether executives have the authority and resources required to deliver. They set the organisation's tolerance for risk, challenge whether the operating model can execute the strategy and assess whether the leadership team has the capability required for what comes next.
The board does not need to run the transformation.
It does need to create the conditions in which the people responsible for delivering it have a realistic chance of succeeding.
Related analysis
We've invested heavily in AI but aren't seeing enough business value. What kind of leadership do we actually need?
How to identify whether the constraint sits in business ownership, data, technology, operating model or executive capability.
Can your CIO lead AI, or do you need a different leadership model?
How to assess whether the existing CIO has the capability, remit and authority to lead enterprise AI.
What do AI and data leaders need from the organisation to deliver?
How authority, sponsorship and organisational conditions shape AI execution.
How to Assess CIO and CTO Candidates When You're Not Technical
How boards can assess technology leadership without trying to replicate the candidate's technical expertise.
Reviewing board and executive capability around AI and technology?
Lucent Search works with boards and executive teams to clarify where accountability should sit, what leadership capability the organisation requires and whether existing roles are designed to deliver.
Where a new executive appointment is required, Lucent conducts retained executive search for senior AI, data and technology leaders.
Comments