McKinsey says enterprise AI is finally ‘on the road to ROI’
McKinsey's latest State of AI report claims enterprise AI is "on the road to ROI," yet its own survey data shows the share of organisations reporting a real earnings boost from AI has barely moved in a year. This matters because it confirms a persistent gap between soaring corporate investment and enthusiasm for AI, and the actual financial returns businesses can point to, even as firms continue pouring money into the technology and increasingly expect it to cut jobs.
The survey of 1,719 professionals found 37% attribute at least some earnings impact to AI, roughly unchanged from 2025, while only 6% qualify as "high performers" attributing 5% or more of earnings to AI with significant impact, also flat year-on-year. Adoption is still climbing regardless: 40% of large firms are scaling AI agents (up from 27%), a third are building coding tools in-house rather than buying software, and 80% of individual users report productivity gains, though a fifth cite rising AI costs as a constraint. On jobs, 39% of respondents now expect AI-driven cuts in the coming year, up from 32% in 2025, though McKinsey notes last year's predicted job losses failed to materialise to the extent expected.
- McKinsey: AI investment keeps rising, but earnings impact stays flat
- Only 6% of firms are AI "high performers" seeing real financial gains
- More respondents now expect AI-driven job cuts, up from 32% to 39%
New here? Start with this
Companies have invested heavily in artificial intelligence, or AI, to help with tasks such as writing, analysis and customer service. But using AI more widely does not necessarily mean it has increased a company’s profits: businesses need to weigh any savings or extra sales against the cost of developing and running the tools.
McKinsey is a consultancy that surveys businesses about how they use AI. In its research, “earnings impact” means respondents say AI has contributed to their organisation’s financial results; it does not by itself show how much AI caused those results to change. Companies are also exploring AI agents, which can carry out tasks through several steps, and tools built for their own staff.
The issue matters to workers and businesses because AI may affect both productivity and staffing. Survey expectations about job cuts are not the same as confirmed job losses, and predictions can differ from what later happens.
Both sides, in good faith
The strongest fair case each way — we don't pick a winner.
The case for
Supporters of McKinsey's assessment argue that measuring transformative technology adoption by current ROI levels misses the point of early-stage adoption cycles. The doubling of AI agent scaling (40% from 27%), widespread individual productivity gains, and the build-out of internal capabilities represent genuine progress that establishes foundations for future returns. Previous predictions of job losses failing to materialise as expected suggest the transition may be smoother than feared. "On the road to ROI" accurately describes this trajectory—modest by absolute measures but genuinely improving and directionally sound.
The case against
Critics contend that McKinsey is reframing stagnation as progress to maintain corporate enthusiasm. The core metrics show no meaningful improvement year-on-year (37% reporting earnings impact, 6% as high performers), yet the company presents this flat performance as positive whilst adoption climbs and job-cut expectations rise to 39%. This framing obscures a troubling disconnect: massive corporate investment without proportional financial returns, now paired with expectations of significant workforce reductions. The narrative appears designed to hide an inconvenient reality rather than acknowledge that extensive spending has yet to deliver measurable financial benefits.