Spanish Point Technologies’ Daire Cunningham explores the human element that’s key to digital transformation, especially in relation to AI adoption.
Artificial intelligence has become remarkably easy to buy, yet surprisingly difficult to transform a business with.
That distinction matters. McKinsey’s latest global research found that 88pc of organisations are now using AI in at least one business function, yet only around a third have begun scaling their AI programmes across the enterprise.
The gap between those figures points to a challenge that is becoming increasingly apparent. Access to AI is no longer the primary obstacle for most organisations. The harder task is understanding where it can genuinely create value and what needs to change within the business for that value to be realised.
In the rush to adopt AI, organisations can easily begin with the wrong question: Which AI tool should we use?
A better starting point is to ask what problem the organisation is actually trying to solve. Once that is clear, leaders can work backwards. They can identify what information is required, where it sits, who should have access to it, how the existing process operates and where the real bottlenecks lie.
They can also determine which decisions should remain firmly in human hands and where technology could meaningfully improve the outcome. The choice of model or platform should follow those decisions, rather than lead them.
This matters because one of the biggest obstacles to AI transformation is not a shortage of sophisticated technology, but the condition of the organisation underneath it.
A 2026 Dun & Bradstreet survey found that while more than three-quarters of businesses are already reporting some measurable return from AI, just 6pc say their enterprise data is fully ready to support AI at scale.
Businesses have spent decades accumulating information across databases, applications, document repositories and departmental systems. Permissions have evolved over time, processes have been layered on top of other processes, and employees have developed workarounds to get things done. AI has a tendency to make the weaknesses in those arrangements much harder to ignore.
Consider an AI assistant connected to an organisation’s internal information. If an employee asks it a question and receives a confidential document they should never have been able to access, it is tempting to describe that as an AI security problem. In reality, the technology may not have created the underlying vulnerability at all. It may simply have exposed the fact that the employee already had inappropriate access.
As AI makes it dramatically easier to retrieve and connect information, identity, permissions and governance become more important, not less.
‘Technology itself is only part of the story’
We have seen the value of getting those foundations right in identity modernisation projects. At the Royal College of Physicians of Ireland, 13 separate systems were brought into a single identity platform, introducing single sign-on and automated access management for more than 86,000 users across 95 countries.
Administrative overhead was reduced by 50pc. The technology itself is only part of the story. Establishing a clearer way of managing identities, permissions and access provides a much stronger foundation for any organisation introducing technologies capable of searching, connecting and retrieving information across multiple systems.
There is an equally important human dimension to this. I have always thought about digital transformation partly in terms of arming the workforce to adopt and employ modern technology in their working practices. An organisation can deploy an exceptional system, but if employees do not understand where it fits into their jobs, why they should use it or how their responsibilities change around it, its value will remain limited.
McKinsey’s research has found that redesigning workflows is among the factors most strongly associated with organisations seeing financial impact from generative AI. That makes sense: transformation does not happen because somebody has been given access to another piece of software. It happens when organisations rethink how work gets done.
At the same time, AI does not remove the need for professional discipline.
Engineers can use it to accelerate coding, testing and analysis, but generated code still has to be reviewed, systems still have to be tested and security requirements still apply. Someone remains accountable for what ultimately goes into production.
The same principle applies across other professions. AI can accelerate elements of the work, but capable professionals still need to understand the outputs, recognise when something is wrong and exercise judgement over the decisions that follow. As systems become more capable, that human oversight becomes more important rather than redundant.
This is also why the strongest applications of AI tend to make more sense when described in terms of the problem being solved rather than the technology being deployed.
The music industry provides a useful example. Identifying potentially fraudulent music registrations across immense volumes of metadata is a specific problem that becomes increasingly difficult to address at scale.
Applying machine learning and AI-driven metadata analysis to that challenge has enabled suspicious registrations to be identified across large datasets, with processing times reduced by 70pc in one implementation. The significance is not simply that the same work can be completed faster. Greater processing capacity allows skilled people to focus their attention on the cases that genuinely require investigation, judgement and expertise.
That points to a broader question about how businesses measure the value of AI. Much of the discussion still focuses on automation, efficiency and ultimately how much human work can be removed from a process.
Efficiency matters, but it is only one measure of what these technologies can contribute. AI can also make previously impractical work possible by analysing volumes of information that would have required enormous amounts of human time, identifying patterns that warrant further investigation and allowing skilled employees to concentrate on more complex decisions.
Instead of asking only what work AI can remove, businesses should also be asking what their people can now do that they could not realistically do before.
There will always be another model, platform or AI capability arriving on the market, which is precisely why businesses cannot build a transformation strategy around chasing each new development.
They need something more durable: a clear understanding of the problems they want to solve, trusted information, appropriate governance, well-designed processes and employees who understand how to use the technology responsibly.
The organisations that succeed with AI will not necessarily be those that adopt every new tool first. They will be those that know where AI belongs, prepare the foundations beneath it and give capable people the structures and confidence to use it effectively.
Technology certainly enables transformation, but people still have to make it work.
Daire Cunningham is the COO at Spanish Point Technologies, where he has worked since 2008. Prior to joining Spanish Point Technologies, Cunningham spent some time working in software systems design and development roles at companies such as Dolmen Stockbrokers and Fyntel.
Don’t miss out on the knowledge you need to succeed. Sign up for the Daily Brief, Silicon Republic’s digest of need-to-know sci-tech news.
You must be logged in to post a comment Login