The AI job displacement story has been hypothetical for three years. In July 2026, Bloomberg published the data that makes it concrete: the financial-activities and information sectors are losing 28,000 jobs per month on average, and AI adoption is the primary cited reason.

So far in 2026, roughly 102,000 announced job cuts have been explicitly attributed to AI. That number is growing, and the industries where it is happening first tell you something specific about where AI automation is most immediately effective.

Which Roles Are Going First

In finance, the pattern is clearest. Office and administrative support occupations, including customer service representatives, bank tellers, and insurance claims processors, account for roughly a quarter of employment in financial services. These are roles with structured, repeatable tasks that AI handles well: processing a claim, answering a balance inquiry, routing a customer to the right team. The tools that do these jobs are cheaper to run at scale than the humans who did them before, and the output quality is measurable enough that organisations are comfortable with the comparison.

In tech, the cuts are hitting customer support, content moderation, data entry, QA testing, and parts of traditional software engineering. Some of these roles have been automated for years in theory. The difference now is that the tools are good enough and cheap enough that organisations are actually making the call.

The company-level examples are instructive:

Oracle eliminated 21,000 positions, roughly 13% of its workforce, citing AI adoption. Atlassian cut 1,600 jobs (10%) to redirect investment toward AI and enterprise sales. Intuit eliminated 3,000 roles, 17% of its total workforce, to reduce complexity and redirect resources toward AI infrastructure. IBM replaced approximately 200 HR positions with AI agents.

The savings from these cuts are not disappearing into margins. They are being reinvested into AI data centres, chips, and tooling. The workforce is shrinking in some areas and the investment in AI infrastructure is growing simultaneously. That is not companies hedging. That is companies executing a strategy.

What the Pattern Actually Shows

It is worth being precise about what this data does and does not mean.

The 28,000/month figure represents jobs in two sectors that happen to have the highest AI adoption rates. It does not mean every sector is experiencing cuts at this rate. Healthcare, construction, education, and trades are not in this data. The AI automation story is sector-specific in ways that get lost when headlines talk about AI jobs displacement as a monolithic trend.

The roles being eliminated are also specific. These are not primarily jobs that require judgment, relationships, or context that changes across situations. They are jobs that require consistency and volume. Processing 500 claims per day in the same way every time is exactly what AI is good at. Managing a client relationship through a complex negotiation is not.

The data also does not tell you whether the workers being displaced are finding comparable work elsewhere. Bloomberg’s job loss data tracks cuts, not outcomes. What happens to workers whose roles are automated is a different and more complex question than how many roles are cut.

The Business Decision Underneath the Headlines

For business owners, this data matters for a practical reason that is separate from the ethics or politics of AI automation.

If your competitors in finance or tech are cutting 10-17% of their workforce and reinvesting that money into AI infrastructure, their cost structure is changing. The businesses that are doing this will, over the next 18-36 months, be able to compete at prices that are harder to match for companies that have not made the same move. That is not true in every industry or every function, but it is true in the specific roles where AI automation is already reliable enough to deploy at scale.

The actionable question is not whether to automate in principle. It is which specific functions in your business look like the roles being automated in finance and tech, and what it would actually cost and save to automate them.

Administrative support, claims processing, customer service routing, data entry, QA checking, and similar structured, high-volume tasks are the clearest current answers. If your business has significant headcount in those functions and you have not started evaluating AI alternatives, the window where you can make that move proactively, rather than reactively, is shrinking.

What This Means for People Building Data Careers

For data professionals, the picture is more nuanced than the headlines suggest.

The roles being cut are not data roles. They are roles that feed data into systems or action outputs of systems. The people building the AI systems, maintaining the data pipelines, evaluating AI output quality, and helping organisations decide which processes to automate are not in the 28,000/month cut cohort. They are in the hiring cohort.

That distinction matters if you are thinking about where to invest learning time. SQL, Python, Power BI, and the ability to translate between business questions and data systems are not becoming less relevant because AI is cutting bank teller roles. They are becoming more relevant, because organisations deploying AI at scale need more people who can manage, evaluate, and improve those systems than they did before.

The skills that protect you in this environment are the skills that sit above the automation layer, not beneath it.

---

Enterprise DNA’s learning platform trains data professionals in the skills that matter most in an AI-first environment. If you are evaluating how to upskill your team for the tools and decisions ahead, the EDNA Learn business plans are built for exactly that.