The question of which jobs AI will take has dominated headlines for three years. The more useful question, and the one this data answers, is which industries face structural transformation, and at what scale. The picture that emerges from Goldman Sachs, McKinsey, and the World Economic Forum’s most recent research is more nuanced than either the alarmists or the dismissers suggest. Some sectors face near-total automation of their routine task base by 2030. Others, despite generating enormous AI coverage, are proving surprisingly resilient. And the single most exposed industry isn’t the one most people would guess.
The chart below ranks 15 major industries by the percentage of their tasks automatable by AI by 2030, alongside estimated global job displacement figures. Use the filters to explore by risk tier. Hover over any bar for an industry-level breakdown.
The most exposed industry isn’t manufacturing; it’s the office
Administrative and office support sits at the top of the ranking with 46% of tasks automatable, ahead of manufacturing at 45%. This surprises most people, because the narrative around automation has historically centred on factory floors and assembly lines. But AI’s core strengths (processing text, managing schedules, routing communications, handling data entry) map almost perfectly onto the daily task list of administrative workers. Office and administrative support roles account for 54% of the top 50 highest-risk occupations identified across 784 roles analysed, with bookkeeping and accounting clerks facing 95% task exposure and medical records specialists at 93%. Globally, an estimated 150 million administrative workers are affected. That is the largest absolute displacement figure of any sector.
Manufacturing’s AI disruption is accelerating, not beginning
At 45% task automation potential, manufacturing sits just behind administrative work, but its displacement story is older and faster-moving. The US alone has lost approximately 1.7 million manufacturing jobs to automation since 2000, with Oxford Economics predicting global manufacturing could lose up to 20 million jobs by 2030 if automation continues to accelerate. What’s changed is the nature of the automation: where previous waves targeted physical assembly, AI-guided robotics are now reaching quality assurance, logistics coordination, and production planning. These are roles that previously required human judgment.
Customer service is the sector where displacement is already happening
With 41% of tasks automatable, customer service sits third in the ranking but it arguably leads on speed of actual displacement. AI chatbots and voice agents are not a future threat in this sector; they are an operational reality deployed at scale right now. Most human customer service interactions are no longer handled by phone with human employees. Queries and problems are increasingly repetitive in nature, and AI provides automated responses to frequently asked questions across the majority of customer touchpoints. Human agents are being retained for complex escalations, complaints, and high-value relationship management, a much smaller slice of total volume.
The white-collar surprise: finance and legal are more exposed than most workers realise
Accounting and financial administration sits at 37% automation potential, and legal research and paralegal work at 30%; both significantly higher than most professionals in those fields would expect. McKinsey’s analysis indicates that 30% of work hours in financial services could be automated by 2030, with investment banks already reporting substantial reductions in junior analyst positions. For legal roles, generative AI can now produce first-draft contracts, research memos, and due diligence summaries at a fraction of the time and cost of a junior associate. The exposure is concentrated at the entry and mid-level; senior judgment, client relationships, and licensed advocacy remain firmly human.
The sectors that are genuinely resilient and why
The bottom three industries in the ranking (IT and software development (20%), construction and trades (18%), and clinical healthcare (17%)) share a common characteristic: they require either licensed human judgment, physical presence and dexterity, or both. Autonomous vehicle technology receives significant attention, but large-scale displacement in construction and trades before 2030 is unlikely in most regions due to the physical complexity of on-site work. Clinical healthcare faces a similar dynamic; AI is transforming diagnostic imaging and administrative workflow, but the licensed judgment, physical examination, and patient relationship at the core of clinical practice remain beyond current AI capability. IT and software development is the most counterintuitive entry in the lower-risk tier: AI coding assistants are dramatically changing how developers work, but the net effect so far has been to increase demand for developers rather than reduce it.
A note on methodology
The task automation percentages in this chart are derived from Goldman Sachs Global Investment Research (2025 update) and McKinsey Global Institute’s Future of Work analysis, which model the share of industry tasks that current or near-future AI systems are capable of performing. Task automation potential is distinct from job displacement: a role where 40% of tasks are automatable may see significant productivity change without headcount reduction, or may see targeted role restructuring rather than outright elimination. The WEF Future of Jobs Report 2025 frames the net global picture as 170 million new roles created against 92 million displaced by 2030 — a net gain of 78 million jobs, though the skills mismatch between displaced and created roles remains the central policy challenge. Global jobs at risk figures represent mid-point estimates across multiple research sources and should be treated as indicative rather than precise. This chart will be updated annually as new data becomes available.
3 thoughts on “Mapped: Which Industries Will Lose the Most Jobs to AI by 2030”