British businesses are spending billions on artificial intelligence systems in the hope they will deliver savings and boost growth. Yet we know that many companies are seeing no measurable return on their investment.
Meanwhile, some of their employees have already figured out how to make AI work, but they have been doing it via the “back-door” route, using systems they’ve subscribed to personally.
The scale of corporate investment is significant. According to data in Santander’s autumn trade barometer, published on Monday, the proportion of businesses planning AI investment has nearly doubled in six months, reaching 40 per cent among companies considering international expansion.
The reason many companies aren’t yet seeing significant returns is not that AI fails to work; it is that workers by and large haven’t been waiting for corporate solutions. Employees are often paying for ChatGPT themselves because they find it useful and accessible; the AI revolution has already begun, but on personal phones and laptops.
The companies extracting real value from AI investment are giving their employees a voice in identifying where automation might genuinely help. They are learning from employees who have already figured out what works and building these lessons into their plans.
Larger businesses find it more difficult to adapt and utilise these types of tools quickly given the need for greater oversight. Our data shows that they are far more likely to have AI investment plans than smaller firms, yet often these plans do not move beyond the pilot stage.
The deeper problem is that businesses are often buying technology without the human capacity to deploy it. When we asked what would most help to build AI capabilities, training and reskilling programmes topped the list, particularly among companies with international expansion aspirations.
Larger businesses, despite leading in AI investment, show no corresponding advantage in workforce development.
With the UK hoping to leverage AI for international competitiveness, this represents a fundamental weakness. Most businesses hoping to expand internationally are confident AI will transform their operations, yet very few believe the UK labour market possesses sufficient skills for companies seeking international growth.
Workers are already demonstrating they can use AI effectively when given access to the right tools. What is missing is investment in scaling that capability.
When discussing the impact of AI on the workforce we should discuss what is happening as well as what may theoretically occur. Redundancies are not yet occurring en masse — although some large corporations have been making a start — but companies successfully deploying AI are seeing selective reductions in customer support and administrative roles, typically in areas that are partially outsourced.
More tellingly, executives in sectors where AI has gained genuine traction now anticipate reduced hiring over the next two years, suggesting the technology will reshape employment more selectively than some headlines suggest, complementing rather than substituting workers.
Government has an obvious role here. A substantial proportion of businesses believe ministers need to support with education and training. A commitment in the government’s trade strategy to digitising export processes may not be effective if businesses lack the skills to capitalise on it. The training programmes and digital infrastructure that businesses cite as priorities require public investment, not just private initiative.
Businesses must recognise that workers are not waiting around for permission to use AI. Leveraging the knowledge of those who understand their job roles inside out and have a good view of what AI can do for them, in conjunction with experienced AI data scientists, will help to bridge the gap between corporate and personal approaches to AI use.
The transformation British businesses anticipate is already emerging from employees who understand their jobs intimately and have seized access to tools that make them more effective. The question is the extent to which management teams choose to learn from them, invest in supporting them and accept that the AI revolution is already well under way.
Frances Haque is chief economist of Santander UK. This is the first in a series of opinion pieces on artificial intelligence