30% of Companies See Productivity Drop After Adopting Agentic AI Tools, McKinsey Finds

A new McKinsey report delivers a surprising finding for teams rushing to deploy AI at scale: in nearly 30% of companies, productivity actually fell after adopting agentic AI tools.

McKinsey's Technology Trends Outlook 2026 pinpoints the cause as a lack of structural discipline. The report calls out "vibe coding" — the trend of using autonomous systems to generate code without systematic oversight — as a primary culprit. More code is not translating into more shipped products: AI tools in one study increased coding activity by 180%, but released features rose by only 30%.

Developer trust is another major friction point. Approximately 46% of developers worldwide actively distrust the accuracy of AI tools, while only 33% express trust, and just 3% highly trust the outputs.

Key Takeaways

• 30% productivity drop: Nearly one-third of companies report lower productivity after adopting agentic AI tools, despite widespread adoption.
• Output vs. outcomes gap: A 180% rise in AI-assisted coding output produced only a 30% increase in shipped releases — more code does not equal more value.
• Developer trust gap: 46% of developers distrust AI tool accuracy; only 3% highly trust it — a meaningful constraint on enterprise adoption.
• Investment surges regardless: Equity investment in agentic software development companies grew from roughly $5 billion in 2025 to over $61 billion in the first half of 2026, driven largely by a $60 billion acquisition of Cursor.

The report warns that organizations are "racing to deploy AI at scale without any proven road maps" and face real challenges around workforce upskilling, legacy systems, and security exposure.

Read the full article on Asianet Newsable