The early signs of an AI-induced economic shift are already visible: investor concern following Anthropic’s launch of a legal plugin for Claude Cowork earlier this year reportedly wiped around £285 billion from global software and professional services stocks in a single day.

This is just one example of an increasing number of scenarios playing out in what analysts are deeming the ‘SaaSpocalypse,” as frontier AI labs target some of the most ubiquitous forms of white-collar work – software engineering, legal services, accounting, data analysis and back-office administration. These are the tasks businesses buy at scale, which makes them the obvious place for AI firms to turn model capability into revenue.

That means the first economic impact in the age of AGI may not feel especially futuristic. It may feel quite ordinary. If a model can replace or compress work that previously needed teams of highly paid knowledge workers, that is a clear productivity gain for the business.

But the remaining value of that work does not disappear; it just moves somewhere else.

Who captures the value?

Competition and margins have long been drivers of economic change. The economic impact of AI will ultimately be shaped by these same dynamics. Once AI becomes good enough at routine professional tasks, competition between model providers is likely to drive down the cost of many knowledge-based services. In some sectors, that could increase demand by making expertise cheaper and more accessible. In others, it could put sustained pressure on margins and fundamentally reshape market sizes.

Either way, where once the value of white-collar work was captured locally by individuals and the businesses that employed them, that value is more likely to be concentrated in a small number of platform businesses and their shareholders.

The effects of that shift could expand into a much larger economic issue. White-collar wages form a major part of household income, and household income is one of the main drivers of consumer demand. If AI materially weakens the earnings base of a very large proportion of knowledge workers, you could end up with a structural demand gap: more output capacity than ever, but fewer consumers with the purchasing power to absorb it.

Most conversations around the economic impact of AI focus on companies losing revenue or individuals losing their jobs. But far less attention has been paid to the second-order effect: what happens to aggregate consumer demand when white-collar income across every industry shrinks?

Adding the geopolitical dimension

As it stands, many countries – including the UK – aren’t well-positioned to retain their redirected white-collar value.

Most frontier AI capability is being built by US and Chinese companies. This leaves British businesses in the difficult position of using and paying for AI infrastructure developed overseas while much of the economic value flows out of our local markets.

If UK firms become dependent on imported AI systems to run their core operations, the productivity gains could be significant. But so could the long-term leakage of revenue, data and strategic control.

This is not simply a commercial issue. Over time, it could shape competitiveness, economic resilience and the extent to which domestic markets retain ownership of the value AI generates.

Keeping the value in the UK

The best way to retain a portion of the value flowing into AI isn’t to try to compete directly with the largest AI labs. Instead, it is to lean into using AI to strengthen the very thing that an individual company already does well.

Most of today’s frontier AI systems are designed to handle broad, general-purpose tasks – that is where the highest return on investment currently exists. But in industries where organisations have proprietary data, regulatory complexity or highly specialised workflows, there is still room to build tailored systems that are very difficult and expensive to replicate.

For example:

• A pharmaceutical company’s peptide design workflow involves mass spectrometry analysis, design of experiments and integration with specific lab instruments and LIMS systems.
• An ecological consultancy quantifying carbon loss from vegetation decline on degraded land needs site-specific remote sensing models, species-level classification and knowledge of the particular carbon accounting methodologies accepted by regulators.
• A quarry operator’s aggregate quality control requires computer vision models trained on site-specific material types and lighting conditions.

Put simply, sector-specific AI models built into proprietary workflows like this may allow businesses to retain more of the value they generate, rather than simply passing it upstream to larger platform providers – at least for the time being.

Recommended reading

• UK Leaders Overestimate AI Readiness, Study Finds
• FCA to Protect Access to Cash Amid Bank Branch Closures
• FinTech Scotland Festival Reveals 2024 Programme

And this principle is bigger than helping any one company succeed; the more businesses that own domain-specific AI, the more distributed the economy remains.

The bigger picture

Ultimately, the fixation on AGI as a technological breakthrough risks obscuring the more important story. The real issue is market structure, labour, ownership and economic power. AGI is simply the catalyst. The central questions are where value accrues, who controls the systems, and which firms are best positioned to capture the gains while defending their margins.

The companies most likely to thrive may not simply be the ones that adopt AI fastest, but the ones that understand where the economic gains are most likely to land – and how to keep a meaningful share of them.