AI can make a task faster without making a business more productive or profitable. Lawyers writing in Financier Worldwide’s December 2026 issue argue that companies risk misleading customers and investors when estimated time savings are presented as proven financial gains.
Watch Desk analysis
What happened
Reuben Vandercruyssen and Lydia Savill of Hogan Lovells Cadwalader say AI claims should account for checking and correction time, output quality, operating costs and whether saved capacity leads to useful additional work. They point to research that complicates the usual speed story: in one study of experienced open-source developers using early-2025 tools, participants later estimated that AI had made them 20 per cent faster, but their tasks took 19 per cent longer when AI was allowed. The researchers have noted that newer systems may produce different results.
The authors also cite a study of 5,172 customer-support agents in which access to a generative AI assistant raised issues resolved per hour by an average of 15 per cent, with uneven gains across workers. Their point is not that AI never helps. It is that one result, or a confident impression, cannot stand in for evidence about a different job or a whole organisation. The Financier Worldwide article sets out the legal and measurement concerns.
Why it matters
A supplier’s claim that a tool cuts drafting time is not the same as evidence that it reduces the cost of delivering a service. If staff spend the apparent saving checking outputs, or the work simply waits in another queue, the headline number may flatter the business case. The authors warn that unsupported claims can create risks in customer sales, advertising and investor communications, depending on what was said and the circumstances.
Their practical test is refreshingly unglamorous: compare AI-assisted work with a baseline, include review and correction, assess accuracy, and track meaningful outcomes over time. “Hours saved” is a start, not a return on investment wearing a little hat.
Our read
The strongest point here is that productivity claims need to survive contact with the work itself. A faster first draft means little if someone must check every line, and a pilot result does not automatically travel to every team. Businesses should measure what changed, what it cost to check, and whether the extra capacity produced anything useful before turning a promising trial into a sales promise.
What to watch
- Whether companies disclose how they measure AI-related time savings and financial returns.
- Whether evaluations include checking, correction and the quality of the finished work.
- Whether regulators or customers challenge claims that move beyond what the evidence supports.
Discussion spark: Should companies have to publish the evidence behind quantified AI productivity claims, or is ordinary advertising and investor scrutiny enough?
Sources and evidence
- AI-washing: when promised productivity outpaces the evidence – Financier Worldwide (11 November 2026, 00:00 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.