Did AI Really Cut Jobs? Only 14% of Companies Say It Did

From Ram's CHANGEReadyX™ from TAPAS Advisory · 00:13:22

Show by Ram Narayanan

Did AI really cut jobs? McKinsey's State of AI 2026 survey — 1,719 respondents across 97 countries — found only 14% of AI-using organisations say AI contributed to a workforce decline in the past year. Less than half the 32% who expected cuts a year earlier. Two-thirds report little or no change in total employment at all. TCS's Enterprise AI Study — 1,333 senior leaders across 25 countries and 12 industries — found the share of leaders who say it is still too early to judge AI's effect on jobs rose from 4% in 2024 to 19% in 2026, and 39% are still assessing whether AI initiatives justify their token and compute costs. In this episode of CHANGEReadyX™ Insights, Ram Narayanan and Priya unpack the layoffs that never came, the certainty that collapsed, and the discipline that stops a leader announcing a cut the AI cannot pay for.McKinsey asked the same question a year apart. In 2025, 32% of respondents at organisations using AI expected AI to reduce headcount over the following year. In the 2026 survey, published 25 August 2026, only 14% said AI had contributed to an overall workforce decline — and two-thirds said AI made little or no difference to total employment. For the coming year, 39% now expect an AI-related decline, up from 32% — a forecast McKinsey's own reading says should be discounted heavily, because the same question overshot the outcome last time. The money tells the same story: in a September Futurum Group survey of 1,636 enterprise technology leaders, 46.9% reported AI spend above plan, and 43.3% of those absorb the overrun. TCS's Enterprise AI Study is the mirror image: 84% rank productivity as the leading AI benefit, yet only 10% let AI agents decide without human approval and only 9% say their agents can securely orchestrate workflows. The debate bank explains why: MIT NANDA found 95% of generative AI pilots show no measurable return on the P&L; S&P Global found 42% of companies abandoned most of their AI initiatives in 2025; Gartner says over 40% of agentic AI projects will be cancelled by end 2027; BCG's 10-20-70 rule says 70% of AI success is people and processes; Workday found roughly 40% of AI time savings lost to rework. The fix is the four-part discipline: diagnose readiness and measure human impact before any headcount decision, redesign the workflow before deciding the headcount, equip managers to carry the change, and measure adoption to value. The episode closes with the diagnostic question: before you announce the next AI layoff, what exactly has the AI earned — and what have you redesigned so it can earn it?Three takeaways: (1) Only 14% report AI actually reduced headcount — McKinsey, 1,719 respondents — against 32% who expected it. Two-thirds saw little or no change. The layoffs were announced; the savings never arrived. (2) Certainty has collapsed: TCS found leaders saying it is too early to judge AI's effect on jobs rose from 4% (2024) to 19% (2026), while 39% are still checking whether AI pays for its own token and compute costs — and 46.9% report AI spend above plan (Futurum Group). (3) Diagnose before you announce the cut — measure human impact, redesign the workflow with a named owner and a measure per workflow, plan the agent like a person (in a position, against an approved budget), and measure adoption to value, not announcements made.Chapters: 00:00 Cold Open — The Layoffs That Never Came · 02:01 What McKinsey Actually Found · 04:16 The TCS Mirror — Less Certain, Still Spending · 06:29 The Pattern — Announced Before It Earned · 09:15 The Four-Part Discipline — Diagnose Before You Announce the Cut · 12:01 The Diagnostic Question — What Did the AI Actually Earn?Run the diagnostic: https://tapasadvisory.com/diagnosticSubstack: https://open.substack.com/pub/ramchangereadyxApple Podcasts: https://podcasts.apple.com/us/podcast/rams-changereadyx-from-tapas-advisory/id6831708010— Ram Narayanan, Transformation Leader | Architect of CHANGEReadyX™, TAPAS Advisory

More episodes of Ram's CHANGEReadyX™ from TAPAS Advisory

More episodes

PodSearch is a free, independent podcast search engine. Read about PodSearch, the privacy policy or contact us.