U.S. healthcare delivery productivity has remained essentially flat for 25 years, even as other service industries made significant gains, but AI could help reverse that trend, according to a Nov. 11 report from McKinsey & Co.
Real gross output per full-time-equivalent employee in healthcare delivery declined approximately 1% between 1998 and 2023, compared with a 56% increase across other service industries, McKinsey found.
The gap has persisted despite decades of technology investments. A separate measure shows a more recent decline: Healthcare delivery productivity fell about 1% annually from 2019 to 2025, while productivity in the broader private economy grew 2% annually, according to Bureau of Labor Statistics data cited in the report.
McKinsey pointed to AI as a potential turning point, citing early evidence that the technology is reducing administrative workloads and generating measurable financial returns.
One study highlighted in the report, published in NEJM Catalyst Innovations in Care Delivery in March 2025, examined ambient AI documentation used by 7,260 physicians across approximately 2.6 million patient encounters. The technology saved an estimated 15,791 hours of documentation time compared with nonusers, equivalent to 1,794 working days.
AI adoption is also accelerating across healthcare organizations. In a 2026 McKinsey survey of 120 healthcare decision-makers, split evenly among payers, providers and healthcare services and technology companies, 92% reported implementing generative AI or pursuing proofs of concept. About 74% said they were implementing or pursuing proofs of concept for multiagent workflows.
However, the financial benefits remain uneven.
Among 111 surveyed leaders whose organizations had implemented generative AI, 38% reported measurable positive returns on investment. Another 38% said their organizations were on track to achieve positive returns but had not yet quantified them.
The 38% reporting measurable returns breaks down as follows: 14% of respondents reported returns of less than twice their investment, 20% reported returns between two and four times their investment, and 4% reported returns exceeding four times their investment.
McKinsey suggested that moving beyond individual AI applications toward redesigned workflows will be important for achieving broader productivity improvements.
The report also found that healthcare organizations are looking to external technology companies to support their AI strategies. Among the 120 surveyed leaders, 48% said they planned to prioritize partnerships for generative AI adoption, while 26% intended to purchase capabilities and 26% planned to build them internally.
McKinsey said that could create opportunities for healthcare technology vendors providing AI infrastructure, data engineering, governance and workflow integration.
McKinsey projects that healthcare services and technology revenue will grow 7% to 8% annually through 2030, reaching more than $500 billion, with AI expected to accelerate growth in software, data analytics and technology-enabled services.
The report also cautioned that not every segment will see a net benefit. AI-driven automation could compress rates for traditional administrative outsourcing businesses that depend on labor-intensive services, slowing their revenue growth by two to three percentage points annually and leaving the segment roughly flat or contracting through 2030.
For providers, McKinsey estimated that productivity improvements and technology-enabled operational redesign could contribute up to 100 basis points of margin improvement. The report noted that its outlook assumes material management action and that its projections outside healthcare services and technology do not yet explicitly reflect AI-driven productivity effects.
That potential upside comes as hospitals face headwinds from ACA subsidy expiration and broader federal policy changes, which McKinsey projects could reduce hospital EBITDA margins by about 80 basis points through 2027.
Despite early signs of progress, the report emphasized that wider productivity improvements are not guaranteed and that healthcare organizations will need to translate AI adoption into operational changes to capture financial benefits.
