AI is producing a striking paradox in the labor market: the same roles seeing the most cuts are also seeing the most hiring. A new global survey of 2,050 executives by Snowflake finds IT operations, software development, cybersecurity, and data analytics experiencing simultaneous contraction and expansion as automation trims repetitive tasks while companies staff up for higher-skill AI oversight.
The split is sharp. Among respondents, 40% report reductions in IT operations due to automation, yet 56% say they are adding headcount in the same function. Software development shows a similar push-pull, with 26% citing cuts and 37% increasing hiring. Cybersecurity posts 25% reporting losses and 46% reporting gains. Data analytics lands exactly in the middle, with 37% noting reductions and 37% noting growth.
Why Losses And Gains Hit The Same Roles And Jobs
This is what technological shifts often look like up close: tasks vanish while roles evolve. Routine work in operations and coding is increasingly automated, but running AI at scale demands new competencies—robust data pipelines, governance, model evaluation, observability, and risk management. Snowflake’s research underscores this transition, with 77% of organizations reporting some job creation tied to generative AI and 35% naming skill gaps as a major barrier to success.
As pilot projects turn into production systems, companies need people who can keep AI reliable, compliant, and cost-effective. That means demand for data engineers, AI operations specialists, and security professionals who understand how to harden and monitor large language models. It also explains why organizations further along in adoption are more likely to see a net positive employment effect: once AI becomes core infrastructure, specialized staffing follows.
Where The Cuts Are Clearer Outside Core IT Functions
Outside IT, the picture is more one-directional. Customer service shows the steepest contraction, with 37% of surveyed organizations shrinking their workforce and only 15% hiring. This reflects the rapid maturation of AI-powered chat and voice agents, combined in some cases with continued outsourcing.
The swings are milder elsewhere. In manufacturing and supply-chain operations, 6% report cuts while 13% are hiring, suggesting targeted automation alongside selective investment in digital and AI-enabled logistics. Marketing remains mixed, with 16% reporting reductions and 12% adding roles as generative tools absorb lower-level content tasks but increase demand for strategy, brand safety, and analytics.
Skills Now In Demand For Reliable, Scalable AI
The survey highlights the technical friction points holding back agentic AI: 42% cite interoperability challenges across tools and platforms, 39% struggle with legacy system integration, and 42% point to the need for real-time data to support AI decision-making. Concerns are not only technical—29% flag job displacement, 29% emphasize the importance of human oversight to prevent rogue actions, and 29% raise issues around data storage and use.
Market signals mirror these needs. LinkedIn reports brisk growth in roles such as AI engineer, data engineer, and AI product manager, while the OECD and World Economic Forum have repeatedly found that AI tends to reshape tasks within occupations more than eliminate entire categories outright. In other words, employability is tilting toward people who can connect models to business outcomes, safeguard data, and ensure systems behave as intended.
How Companies And Workers Can Respond To AI Change
Enterprises should assume the paradox persists and plan for talent reallocation rather than pure headcount reduction. That means building LLMOps discipline, funding data quality programs, and creating clear governance for model risk. Several large employers have moved early: Accenture has announced multibillion-dollar AI investments and training for hundreds of thousands of employees, and Amazon’s AI education initiatives aim to broaden foundational skills across the workforce.
For workers, the safest path is to steer toward oversight and integration. Upskill in cloud data platforms, feature engineering, prompt design with guardrails, model evaluation techniques, and cybersecurity for AI systems. Professionals in QA, analytics, and operations can pivot into roles testing model accuracy, tracing data lineage, and maintaining performance SLAs for AI services—areas less likely to be automated and more likely to grow.
The Bottom Line On AI’s Impact On Tech Employment
AI is compressing low-level tasks while inflating demand for higher-order skills in the very same jobs. The result is a job market that can shed and hire simultaneously, often within a single team. Organizations that invest in data foundations and human oversight will capture the gains; individuals who learn to steward AI—rather than compete with it—will have the wind at their backs.
