The rapid development of Artificial Intelligence has dominated financial news for several years, and investors have agonised over a deceptively simple question: will AI prove genuinely transformative, or will it follow the well-worn path of innovations that promised to reshape the world and quietly failed?

For UK investors, answering that question means looking past the headlines about American chip makers and Silicon Valley giants to understand where disruption is most likely to land, and crucially, where it already has.

The answer is where you least expect it.

Just last week, markets were rattled not by Nvidia or Microsoft, but by tremors in financial services. US insurance brokers fell sharply when Insurify unveiled a ChatGPT-powered application capable of comparing car insurance across the entire market in moments. Wealth management stocks followed when Altruist launched Hazel, a tool helping financial advisers build personalised client strategies at a fraction of the traditional cost. For British investors with exposure to insurance and wealth management (and many will have it, given financial services’ weight in the FTSE), these developments deserve close attention. What happens in the United States today has a reliable habit of arriving here shortly afterwards.

Yet it would be a mistake to read these events as confirmation that AI will simply hollow out entire professions. What Insurify and Hazel are doing is streamlining the processing layers of financial services: comparison, aggregation, template personalisation, not replacing the accumulated judgement of an experienced underwriter or a seasoned investment manager navigating clients through volatility. The intellectual capital at the top of these professions is considerably harder to automate than the workflows surrounding it. The more likely outcome is a fundamental reshaping of how these industries are staffed, with significant consequences for middle and back-office headcount, rather than the wholesale replacement of professional expertise.

This analysis becomes particularly interesting when considering the fate of established Software-as-a-Service businesses, until recently regarded as among the most dependable compounders in technology. Companies like Salesforce, ServiceNow, and Sage built formidable businesses on a straightforward model: charge recurring licence fees for deeply embedded software and rely on the friction of switching to protect revenue year after year. These moats were the envy of the investment world.

AI is now testing them in ways not anticipated, even two years ago. Salesforce has moved aggressively to integrate AI through its Agentforce product, promising to automate customer service and sales tasks that previously required human intervention. The strategic logic is clear. Embed AI or a nimbler competitor will use it to undercut you. But this creates an uncomfortable dynamic. The efficiency AI delivers, tends to compress the number of software seats a business needs. If an AI agent can do the work of five customer service representatives, the customer may need fewer licences, not more. Salesforce is attempting to raise the value of each transaction even as AI risks reducing transaction volumes across the industry. Whether that trade-off works in its favour remains genuinely uncertain.

The broader SaaS model faces a more fundamental challenge. Much of what these businesses charge for (i.e. workflow automation, data management, process standardisation), is precisely what AI does best. New entrants are already building AI-native alternatives at significantly lower price points, without the legacy architecture that makes incumbent systems both sticky and, increasingly, cumbersome. For UK investors holding global technology funds with meaningful SaaS exposure, this is not a distant theoretical risk. It is arriving now.

As with every technology transition, there will be winners and losers. The analogy that matters most is not the DeLorean, a product that failed through poor execution rather than a flawed concept, but the shift from packaged software to cloud computing fifteen years ago. That transition destroyed some businesses, transformed others, and created entirely new categories of winner. Microsoft, written off repeatedly during that period, emerged stronger because it owned foundational intellectual property the new world still needed. The lesson is the same today: those who own the underlying models, the data infrastructure, and the core IP on which AI applications are built will collect growing royalty streams regardless of which specific applications ultimately prevail.

For UK investors, the practical implication is to look beyond the obvious beneficiaries and ask harder questions. Which businesses are genuinely building with AI, which are merely bolting it on, and which are quietly most exposed to its disruptive reach. Salesforce’s share price, down over 30%, is a timely reminder that markets are not euphoric right now, so much as grappling with how this technological shift will actually play out over the years ahead.