LO
Business analyst — tasks, one by one
The unit of analysis is the task, not the job title. Each one below carries its direction, whether the judgement rests on evidence or on platform inference, the reasoning, and what it does not establish.
Every task on this page#
Eliciting requirements
Being augmented≈ Platform inferenceInterviewing stakeholders and users to find out what they need, including what they do not think to say.
A chatbot can now run a structured requirements interview: in 33 simulated interviews, one made a similar number of common mistakes as a human interviewer and elicited up to 73.7% of all requirements. Its authors still consider human-led interviews necessary for sensitive or complex elicitation, where interpersonal dynamics uncover unspoken requirements. US projections expect employment in the occupations that include business analysts to grow.
Simulated interviews with students as stakeholders; nothing here measures real projects or analysts' hours.
Writing requirements and user stories
Being augmented✓ Evidence-backedTurning what was learned into requirement documents, specifications and user stories that developers and testers can work from.
This is where language models have arrived. The US Department of Veterans Affairs lists a deployed system that generates user stories and summarises workshop outcomes, and the Social Security Administration one that generates business requirements documentation from legacy code; a study found GPT-4 drafts of a requirements specification comparable to an entry-level engineer's, in a fraction of the time. In real settings the gain is smaller: at an IT consultancy the analyst put the saving at 10–15% and found the drafts missed knowledge that was never written down, and at a postal group's IT teams an agent that rewrote user stories needed the product owner to validate its output.
Inventory entries describe systems, not their effect on staff; the studies use a university project, one consulting project and a small pilot.
Mapping and analysing processes
Being augmented≈ Platform inferenceMapping how work is done today, finding where it breaks, and designing how it should work after the change.
Software can reconstruct a process from system logs and draft diagrams, but deciding what the process should become involves the people who run it and trade-offs someone answers for. No primary source here measures how much of this work tools now do.
An inference from how the work is described; no primary source recorded here measures process-mapping work or tools' share of it.
Workshops and stakeholder alignment
Still human-led≈ Platform inferenceRunning workshops, resolving conflicting demands between departments, and getting agreement on scope and priorities.
Tools can summarise what a workshop concluded, but getting departments to agree is negotiation between people. US projections expect management analysts' employment to grow 10 percent from 2025 to 2035, and computer systems analysts', 8 percent; O*NET files business analyst job titles, such as IT business analyst, under both.
The projections cover wider occupations that include consultants and systems analysts, and count jobs rather than tasks.
Acceptance and change management
Still human-led≈ Platform inferenceChecking with users that what was delivered meets the requirements, handling change requests, and preparing people for the new way of working.
Acceptance is a judgement made on behalf of the people who will use the system, and in the pilot recorded here even AI-improved user stories were validated by a person before use.
No primary source recorded here measures acceptance or change-management work; it is an inference.