An independent study reference written by Dr Phuc V. Nguyen. It is not official subject material — for assessment requirements always follow your subject outline and vUWS.
How the analyst role is changing
Automation does not remove jobs whole, it removes tasks. Autor (2015) sets out the mechanism: where a machine substitutes for a task, the value of doing that task by hand falls, and where a machine complements a task, the remaining human tasks become more valuable and more in demand. For analysts, the substituted tasks are mechanical ones such as writing routine queries, reformatting extracts and assembling a standard chart. What is complemented is everything the machine cannot check for itself: whether the question is the right one, whether the data can bear the claim, whether the relationship is causal, and whether anyone acts on the result.
Why it matters
Spreadsheets did not end the accountant. They ended the part of accounting that was arithmetic by hand, and the profession moved up into interpretation and control. Analytics tooling is doing the same to query writing and chart assembly. If your value is that you can produce the artefact, that value is falling. If your value is knowing which artefact is worth producing and whether to believe it, that value is rising.
A tool now writes most of the routine queries a team used to write by hand. On the substitution and complementarity logic, which analyst skill should rise most in value?
Worked examples
An analyst asks a generative tool for the top ten drivers of churn. It returns a ranked list with confident explanations, and the top entry is "contacted support in the last 30 days".
The output is a correlation ranking, and the confident wording invites a causal reading. Contacting support is largely what people do when they are already thinking about leaving, so acting on it by making support harder to reach would be exactly backwards. The tool did the mechanical part correctly and had no way to know which direction the arrow points. Judging what the ranking can and cannot support is the analyst's work, and it is now the more valuable half of the task.
A team automates its routine reporting and the reporting analyst's recurring workload halves. The manager asks what the role should become.
Redirect the freed time to the parts nobody automated. Someone has to check that the automated reports still mean what they meant when a source system changes, which is monitoring work that did not exist before. Someone has to sit with the operations team, find out which of the twelve reports actually changes a decision, and retire the rest. Someone has to take the harder questions the team never had time for. The role becomes narrower in production and wider in judgement.
Common mistakes
- ✗Automation will replace analysts. It replaces tasks. Routine production of queries and standard charts is being substituted, while framing a problem, judging what the data supports and getting a decision made are being complemented and are becoming harder to avoid.
- ✗The safe response is to learn whichever tool is currently ascendant. Tools turn over faster than careers do. What survives a tool change is the reasoning underneath: how to structure a problem, what the data can support, and how to explain a result to someone who has to act on it.
- ✗If a tool produces an answer quickly, the analyst's job is finished. Cheap production has moved the bottleneck to verification. The scarce work is deciding whether the answer is trustworthy and whether the question was worth asking in the first place.
- ✗Automating a task lowers the skill the job requires. Automating the routine leaves people with the exceptions, which are the hard cases, plus the job of monitoring a system that is usually right. Bainbridge (1983) named this irony: the residual work is harder than the work removed, and the daily practice that used to build the skill has gone.
Revision bullets
- •Automation removes tasks, not whole jobs
- •Substituted: routine queries, reformatting, standard chart assembly
- •Complemented: framing, data judgement, causal caution, communication
- •The bottleneck moves from production to verification
- •Ironies of automation: the residual work is the hard part and practice disappears
- •Learn the reasoning, not the tool of the moment
Quick check
A tool now writes most of the routine queries a team used to write by hand. On the substitution and complementarity logic, which analyst skill should rise most in value?
A team automates its monthly reporting and disbands the manual checking routine, on the reasoning that the automated version is more reliable than people were. Two years later a source system changes a field and the error goes unnoticed for four months. What does this illustrate?
Connected topics
More in The Analyst and What Comes Next
Sources
- Autor, D. H. "Why Are There Still So Many Jobs? The History and Future of Workplace Automation." Journal of Economic Perspectives, 29(3), 3-30, 2015.The substitution and complementarity argument, and the point that tasks requiring tacit judgement resist automation even when adjacent tasks do not.
- Bainbridge, L. "Ironies of Automation." Automatica, 19(6), 775-779, 1983.Automating the routine leaves the operator with the exceptions and with a monitoring task, while removing the practice that built the skill to handle them.
- Krakowski, S. "Human-AI agency in the age of generative AI." Information and Organization, 35(1), 100560, 2025.Treats agency as distributed by design between people and generative systems. The reading that human work moves towards direction-setting and evaluation as tools take over more of the production step is this atlas's synthesis, not a quoted claim.