Computer Systems Analyst

Computer Systems Analyst scores 16/100 on this site's structural exposure measure, which averages two independent task-exposure studies: OpenAI's "GPTs are GPTs" ratings and the AIOE index. That is a low safety reading, and the usage data points the same way. Microsoft's analysis of Bing Copilot conversations gives analysts an AI applicability score of 0. 31, higher than about 90% of the 785 occupations it covers. The Anthropic Economic Index, which measures observed Claude use, puts the occupation at 0.

28, above roughly nine in ten of the 756 occupations it tracks. Two products and two userbases agree that this kind of work is already being done with AI help.

The shape of the work explains why. Writing requirements, drafting user stories and acceptance criteria, producing test cases, documenting a process and querying a database in SQL are text-and-structure tasks, which is where language models are fastest. Atlassian, whose Confluence is on O*NET's list of analyst tools, offers a Rovo agent aimed at product teams whose example prompts include generating a product requirements document and suggesting user stories. Nothing legal protects the task either.

BLS's how-to-become section names no license or certification for this occupation, and we found no rule that requires a human to sign a requirements document, so a software tool can do the drafting under a manager's or developer's account instead.

The picture is not one-sided, and the page would mislead without the other half. BLS projects 8% employment growth for 2025 to 2035, labelled much faster than average, with about 32,900 openings a year. Its outlook text says that as organizations expand their reliance on IT, including AI, analysts will be hired to design and install new systems. Compare computer programmers, where BLS projects a 7% decline and expects companies to use AI to automate repetitive programming.

We searched for a named employer that replaced systems or business analysts with AI and found none. What the evidence shows is highly exposed tasks alongside a growing headcount projection, not measured displacement.

Where the pressure is most likely to land is the content of the job and its entry rung. Employer-reported layoff data is not specific to analysts: Challenger, Gray & Christmas counted 155,126 announced technology-sector cuts through August 2026, up 52% on a year earlier, and 116,175 cuts across all industries that employers attributed to AI (its September report, released in October, raised these to 165,925 and 120,136). Those are announcements, not verified causes.

The nearest occupation-level evidence is Stanford's payroll study of software developers, where workers aged 22 to 25 show large declines while older groups mostly grow. It does not cover analysts, and its authors describe a correlation, not a cause. Our reading is that if drafting and documentation shrink, the analysts who remain will be the ones who run the workshops, settle disputes between departments and own the outcome.

Two limits apply to every number above. Both structural inputs rate task content, not hiring. The Microsoft conversations date from January to September 2024, so tools released since are not reflected. A single reader's job can sit well above or below the occupation average.