Opinion

McKinsey's 2026 survey: AI revenue gains cluster in marketing and sales, and small companies aren't catching up

The State of AI 2026 puts agent scaling at 40% inside billion-dollar enterprises and 22% at smaller organizations — a widening gap in the exact functions where small businesses compete for customers.

By Aaron Gerstle· Opinion & analysis·2026-09-04·3 min read

Photo: Unsplash / Campaign Creators — Marketing and sales team reviewing analytics on a laptop in a small office

McKinsey’s State of AI: Global Survey 2026, released August 25 and drawn from 1,719 respondents across 97 nations fielded between May 4 and June 8, reports that 40% of billion-dollar organizations are now scaling AI agents, up from 27% the year before. At smaller organizations the figure is 22%, and it hasn’t moved. The same split shows up one layer deeper: 54% of large-org respondents say they’re scaling AI across the enterprise, versus 22% at smaller companies.

Where’s the money showing up? Revenue gains from AI are most commonly reported in marketing and sales, strategy and corporate finance, and product and service development. Those aren’t back-office wins. They’re the functions where a 12-person company competes head-to-head with a 12,000-person one for the same customer.

The overall payoff remains modest. Some 37% of respondents “attribute at least some EBIT [earnings before interest and taxes] impact to AI use,” which the report calls “about the same” as 2025. Only 6% qualify as McKinsey “high performers,” meaning at least 5% of EBIT attributed to AI and impact described as “significant.” As the report puts it: “Organizations’ conviction in AI is growing faster than the immediate financial returns they can attribute to it.”

Michael Chui, the QuantumBlack senior fellow who coauthored the report, told The Register the high performers are seeing “real ROI,” adding, “It’s a journey, not a destination.”

The other structural signal is build-versus-buy. Some 32% of organizations have decided against off-the-shelf software, opting instead to build with agentic coding tools, and the sector spread is wide: 41% in technology, 39% in healthcare payers and providers, 38% in professional services and in energy and materials, but only 19% in insurance and 17% in the public and social sector. Nearly half of high performers skip software purchases, against 31% of everyone else.

“Leaders are asking what their organisations need to build AI tools themselves. The rise of software coding agents and in-house development is one clear sign of this broader shift,” said McKinsey senior partner Lieven Van der Veken.

One in five respondents say their organization is limiting AI use because of operating costs. And 39% expect their employer to cut jobs due to AI in the coming year, up from 32% in 2025.

The pattern rhymes with the ERP wave of the late 1990s, when SAP and Oracle deployments moved enterprise productivity forward and left mid-market firms a decade behind on the same workflows. The revenue-generating layer, Salesforce’s Agentforce crossed $1.5 billion in ARR even as OpenAI courted small businesses directly, is where the gap is compounding fastest. Conviction isn’t the constraint. Distribution is.

Sources