32% of Organizations Skipped a Software Purchase Because Coding Agents Could Build It — but in the Same McKinsey Survey, AI's Financial Return Did Not Move

McKinsey's State of AI 2026 survey covers 1,719 respondents across 97 nations, fielded May 4 to June 8 and published August 25. One finding: 32% say their organization decided against purchasing at least one software product or feature because the functionality could be built in-house with agentic coding tools. Industry variation is wide — technology firms at 41%, healthcare payers and providers 39%, professional services and energy and materials 38% each, insurance at 19% and the public and social sector at 17% at the bottom. Among the 6% McKinsey classifies as high performers (attributing at least 5% of EBIT to AI), the figure approaches 50%. Yet in the same survey, the share reporting enterprise-level financial impact is unchanged from last year — 37% attribute some EBIT impact to AI.

What This Number Says, and What It Does Not

Start with the question as asked: respondents were asked whether their organization had forgone purchasing a software product or feature because the functionality could be built internally with agentic coding tools. 32% said yes. What they were not asked: whether the replacement got built, shipped, passed an audit, or is still running a year later. So the accurate reading is "a decision not to buy," not "a system running in production." That distinction matters this year in particular, because the same report puts the share attributing at least some EBIT impact to AI at 37% — flat against last year, unmoved. Purchasing decisions, in other words, are running ahead of financial results. That is not itself surprising — avoided spend takes a fiscal year or two to show up — but it does mean treating 32% as evidence that SaaS is being displaced is premature.

The Industry Split Carries More Information Than the Headline

Technology firms lead at 41%, then healthcare payers and providers at 39%, professional services and energy and materials at 38% each, financial institutions at 36%, media and telecom at 34%, pharma at 33%. At the other end, insurance sits at 19% and the public and social sector at 17%. The distribution is intuitive: the stronger the in-house engineering bench and the more idiosyncratic the business logic, the greater the pull toward building; heavily regulated sectors with high change costs go the other way. For teams selling B2B SaaS, the number to look at is not 32% but where your own industry lands — the spread is more than twofold. The high performer group diverges further. McKinsey defines high performers as the 6% of respondents attributing at least 5% of EBIT to AI, and they skip purchases at nearly 50% against 31% for everyone else. They are also twice as likely to have scaled software coding agents, and 2.7 times more likely to have scaled other agentic AI.

The Numbers to Read Alongside It

McKinsey's sample is 1,719 respondents across 97 nations, fielded May 4 to June 8, 2026 and published August 25 — roughly three months old by now. Separately, about 20% of respondents report that AI-related operating costs, including token costs, constrained their AI use, though most plan to increase investment. There is both corroboration and counter-evidence. Retool's 2026 Build vs. Buy report says 35% of enterprises have already replaced SaaS with custom software and 78% expect to build more internal tools in 2026, based on 817 Retool customers and builders surveyed in late 2025. But deployment maturity is another matter: Gartner's 2026 CIO Survey found only 17% of organizations have actually deployed agents, and Deloitte's 2026 Tech Trends puts production-ready agentic systems at 11%. Read together, the defensible conclusion is that intent to build has risen sharply while the share actually running is far lower. For teams facing a build-versus-buy decision, the question worth adding is not "can we build this" — with today's tools you probably can — but "who maintains it for the next three years." This site previously covered the NBER survey of nearly 6,000 executives where more than 90% reported no effect on employment or productivity after three years of AI. Both datasets point at the same thing: the lag between input and output is longer than the marketing suggests.

via: McKinsey: The State of AI, Global Survey 2026, Yahoo Finance analysis, Startup Fortune