AI agents threaten to transform white-collar work, offering cheaper, faster software capable of performing tasks once reserved for lawyers, bankers and accountants. But a dozen chief executives, investors and analysts interviewed by the Financial Times say the disruption will not hit every sector equally. In industries where accuracy, accountability and regulation matter, trust is emerging as the critical faultline that artificial intelligence cannot easily cross.
"Fiduciary professions like law, tax, audit and compliance require more than accelerating everyday knowledge work," Steve Hasker, chief executive of Thomson Reuters, told the Financial Times. "What ultimately matters is whether the output is authoritative, traceable and accountable to professional standards." Hasker added that "almost right isn't good enough, whether in a courtroom, a boardroom or an audit," "Hvylya" reports.
The assessment comes after Anthropic released a series of AI tools this year - including Claude Cowork and customizable plug-ins - that allow professionals to automate contract reviews, financial modeling, job application screening and other tasks. The launch triggered a stock market sell-off in February dubbed the "SaaS-pocalypse," indiscriminately hitting software companies from cyber security firms to financial service platforms.
Jean-Baptiste Brian, co-chief executive of private equity firm Hg Capital, said AI plug-ins can improve efficiency and replace some human roles, but cannot fully substitute core enterprise software built on years of accumulated proprietary data. "Plug-ins can't replace proprietary data, systems of records, regulatory context, network effects and distribution - that's my defensibility," Brian said.
Thomson Reuters has already launched CoCounsel Legal, an AI tool powered by Anthropic's latest model, designed to perform complex legal research, document analysis and drafting tasks using proprietary content. Hasker framed the move as proof that models on their own lack the grounding in domain expertise that fiduciary work demands. Victor Englesson, a partner at investment firm EQT Partners, agreed the technology is "very powerful" but said the key question is execution: software groups must prove customers will pay for AI capabilities by showing clear return on investment.
Also read: how one finance graduate switched to manual labor as white-collar jobs dried up across America.