■ HIGH RISK ■ Technology
Yes, in large part — the documentation-heavy version of this job is being automated in real time. What survives is the analyst who owns the decision rather than the deck.
“Gathering requirements and writing specs? AI does this while you're still scheduling the meeting.”
Our AI replacement risk score — how we score jobs
Strip the title back and the classic BA day is translation work: sit with stakeholders, extract what they actually want as opposed to what they said, turn it into user stories, acceptance criteria, process maps and BRDs, chase sign-off, then trace changes through JIRA and Confluence while writing the SQL that answers 'why did conversion drop in the Midwest.' It is a role built on producing artifacts, and artifact production is precisely what large language models are best at.
The automation is not speculative. Meeting transcripts now become structured requirements and draft stories automatically. Natural-language querying against a warehouse means a product manager can pull the cohort analysis without filing a ticket. Process-mining tools reconstruct as-is workflows from system logs faster and more accurately than a month of interviews. Documentation, test cases, gap analyses, and stakeholder summaries all draft themselves to a decent first pass. WEF's Future of Jobs surveys have repeatedly listed data-entry and routine analytical roles among the declining categories, and the artifact-producing BA sits uncomfortably close to that description.
There is a real residue, which is why the score is 72 rather than 90. Requirements elicitation is often political — two departments want incompatible things and someone has to broker it, which means reading the room and having enough institutional credibility to say no. Legacy systems carry undocumented behavior that exists only in the heads of people who have been there eleven years. Regulated domains demand traceable, defensible decisions with a named owner. And the model that drafts a beautiful spec has no idea that the finance team's 'revenue' means something different from the sales team's. The BAs who thrive stop being scribes and become decision-makers: owning outcomes, shaping what gets built, and using the tooling as a drafting engine rather than defending their word count.
Automatability: our editorial assessment of current and near-term AI capability
Already biting. Teams are quietly not backfilling BA roles and folding the work into product managers and engineers who now draft their own specs. Through the late 2020s expect the pure documentation BA to become rare outside heavily regulated industries, where audit trails keep the role alive. By around 2030 the surviving title likely means something closer to product owner or decision analyst, with headcount well below current levels.
Only as a stepping stone rather than a destination. The underlying skills — structuring vague problems, working with data, mediating between technical and business people — remain valuable, but the job title that packages them is contracting. If you enter, aim to move quickly toward product ownership, data analytics with real modeling depth, or a regulated domain specialism.
The documentation and reporting portion is largely automatable today; adoption is the bottleneck, not capability. Most organizations will take several more years to restructure, and regulated sectors longer. Practically, expect the next three to five years to remove much of the junior tier, with senior analysts persisting where accountability and stakeholder politics dominate the work.
Institutional memory and political standing. Knowing that the legacy billing system silently rounds differently for one customer segment, or that a particular VP will block anything touching their team's headcount, is not in any document for a model to read. Analysts who hold that knowledge and can broker decisions between people with competing interests stay valuable.
Product management is the most natural jump — same stakeholder skills, but you own outcomes instead of specs. Data analytics or analytics engineering works if you enjoy the technical side and go deeper than dashboards. Solution architecture and regulatory or compliance analysis are strong options in finance, healthcare, and insurance, where traceability requirements keep humans firmly in the loop.