The firm’s new report, ‘AI productivity gains and the performance paradox,’ concludes that most current AI applications simply accelerate existing work without redesigning workflows. McKinsey aims for a 1:1 parity between its 40,000 human consultants and 40,000 AI agents by year-end.
McKinsey’s strategy practice has released an analysis on the ‘AI paradox’, describing the corporate challenge of integrating AI: while AI adoption and capital investment are increasing, sustained performance impact remains elusive.
The report, ‘AI productivity gains and the performance paradox’, claims that most AI tools accelerate current tasks but maintain underlying workflows. Larger productivity gains will occur when organizations redesign processes around AI instead of merely adding it.
The report’s historical analogy involves electricity in factories. Initially, businesses replaced steam engines with electric motors, gaining efficiency but not reconfiguring layout.
The 💜 of EU tech
The latest rumblings from the EU tech scene, a story from our wise ol’ founder Boris, and some questionable AI art. It’s free, every week, in your inbox. Sign up now!
The true breakthrough came when small motors allowed for a reconfiguration around workflows, and ultimately when companies redesigned factories around electricity, creating new models.
According to McKinsey, general-purpose technologies “rarely create value in a single wave.”
McKinsey suggests executives should: assess AI’s industry impact, build AI-driven competitive moats, and leverage speed into a structural advantage.
The report contrasts companies like JPMorgan Chase, BMW, and Siemens, who accelerate work with AI, versus those who redesign processes to transcend the ‘gen AI paradox.’
The skeptical evidence McKinsey is publishing
With AI investment returns increasingly scrutinized, the Federal Reserve Bank of St. Louis records 1.9% excess cumulative productivity growth since ChatGPT’s launch, despite needed higher rates for current AI capital spending.
JPMorgan’s capex analysis indicates $650 billion annual revenue needed to sustain a 10% return on AI infrastructure, likening it to the late-1990s telecom fibre buildup with unmet revenue expectations.
Research from MIT Media Lab reveals 95% of organizations see no measurable returns from AI adoption. Deloitte reports 66% see productivity gains from AI, but only 20% see revenue growth, with 34% using AI for significant transformation.
PwC’s CEO Survey finds 56% gained nothing from AI investments, while 12% report both revenue growth and cost reductions. Workday’s research indicates 37-40% of supposed AI time savings are used for output correction.
OpenAI president Greg Brockman’s claim that AI writes 80% of its code contrasts a February 2026 NBER finding that 80% using AI report no productivity impact.
McKinsey’s estimation that AI could add $4.4 trillion to the global economy clashes with Nobel laureate Daron Acemoglu’s 0.5% productivity projection. This gap influences enterprise AI funding decisions.
McKinsey’s AI deployment
McKinsey’s skepticism is significant given its own AI integration. CEO