Agentic AI Adoption Hits Productivity Snag in 30% of Firms, Finds McKinsey

Productivity declined at nearly 30 per cent of companies after their teams adopted agentic artificial intelligence tools, according to McKinsey's Technology Trends Outlook 2026. The finding points to operational friction as enterprises race to deploy autonomous coding systems without established playbooks.

The report notes that agentic software development can produce unintended outcomes when adopted without structural discipline. Higher coding output does not necessarily translate into business results: one study cited in the report found that AI tools raised coding activity by 180 per cent, while shipped releases grew by only 30 per cent.

Developer sentiment adds another constraint. About 46 per cent of developers worldwide actively distrust the accuracy of AI tools, compared with 33 per cent who trust them, and only 3 per cent who highly trust their outputs. That trust gap matters for companies seeking to capture the benefits of agentic development, the report observes.

Capital is nonetheless flowing in at speed. Equity investment in agentic software development companies climbed from negligible levels through 2023 to roughly USD 5 billion in 2025, then crossed USD 61 billion in the first half of 2026, largely driven by a USD 60 billion acquisition of Cursor. Job postings in the field rose 221 per cent between 2024 and 2025.

Broader technology spending follows a similar trajectory. AI infrastructure spending doubled in a single year, while energy technologies attracted nearly USD 200 billion in investment during 2025 to meet power needs.

The report cautions that the pace of change brings its own difficulties. Organisations are deploying AI at scale without proven road maps, even as they confront workforces in need of upskilling, legacy systems requiring updates, and networks exposed to evolving security risks.