AI and your next move · 8 min read
Will AI replace data analysts? What the work is becoming
AI can write queries and charts. It cannot decide which question matters or whether the data can answer it. A task audit and one checking exercise for analysts.
The short answer
AI is changing data analysis rather than ending it. Tools can now write queries, clean data and draft charts from a plain-English request, which shrinks routine reporting work. The parts that remain human are framing the right question, knowing whether the data can actually answer it, spotting errors and bias, and persuading people to act. Analysts who grow those skills and use AI to move faster are well placed.
On this page 7 sections
What shrinks and what grows
| Task | Hours a week | AI can draft it? | Who checks and decides? | Keep, shrink or grow |
|---|---|---|---|---|
| Writing routine SQL queries | 6 | Yes, drafts | Me, against known totals | Shrink |
| Weekly dashboard refresh | 4 | Largely automatable | Me, spot-checks | Shrink |
| Clarifying what a stakeholder actually needs | 2 | No | Me | Grow |
| Checking data quality issues | 3 | Can flag patterns | Me | Grow |
| Presenting findings and recommendations | 2 | Drafts slides | Me | Grow |
The skill AI makes more valuable: checking
An AI tool will confidently produce a number. Whether that number is right depends on joins, filters, definitions and data quality that the tool may not understand. Analysts who can reconcile a figure against a known total, spot double counting and explain uncertainty become more valuable as more analysis is generated quickly.
A practice exercise
Take a small public dataset you understand. Ask an AI tool a business question about it in plain English, for example which month had the highest average order value. Then answer the same question yourself in a spreadsheet. Compare. Where they differ, find out why. Common causes are missing values treated as zero, duplicate rows, or a different definition of average.
Write a short note: the question, both answers, the reason for any difference, and what you would tell a manager. That note is the kind of evidence that distinguishes an analyst from a tool operator.
Directions that build on analysis
- Analytics engineering: building reliable data models others depend on.
- Experimentation and product analytics: designing tests and interpreting results.
- Data governance and quality: making data trustworthy.
- Domain specialist roles: health, finance, operations or policy analysis where context is everything.
Questions to ask before trusting any number
- What is the exact definition of each field used?
- Were duplicates removed, and how were missing values handled?
- Does the total reconcile with a figure someone already trusts?
- Has the time period or filter changed since last time?
- Would the conclusion change if one outlier were removed?
These questions are quick to ask and catch most of the errors that slip through generated analysis.
A portfolio that stands out now
Instead of a dashboard alone, show your reasoning: the question, the data's limitations, an AI-assisted first pass, the errors you found and the final answer with its uncertainty. A short write-up like this demonstrates exactly the judgement employers are looking for as routine analysis gets cheaper.
Questions people ask
- Is data analysis still worth learning with AI?
- Yes, but learn it as reasoning, not just tool skills. Understanding data quality, statistics, and how a question maps to data is what lets you check AI-generated analysis.
- What tasks of a data analyst can AI do?
- Drafting SQL, writing spreadsheet formulas, generating charts, summarising results and suggesting cleaning steps. Each still needs checking against the data and the business context.
- What should data analysts learn next?
- Domain knowledge in the area you analyse, experiment design, data quality and governance, communicating uncertainty, and working with AI tools critically.
- Is entry-level data analysis harder to get into now?
- Some early-career evidence suggests young workers in highly AI-exposed occupations have seen weaker employment trends, though researchers caution against treating this as proof of cause. Show judgement in your portfolio, not just output.
Exercises are practice, not evidence of career fit or a qualification. How to read our guides and report a correction.
Sources
- Generative AI and Jobs: A 2025 Update · International Labour Organization
- Canaries in the Coal Mine? Six Facts, revised August 2026 · Stanford Digital Economy Lab
- Data Scientists (Occupational Outlook Handbook) · U.S. Bureau of Labor Statistics
How this page is written
Sources support the factual background, not a guarantee that an exercise predicts career fit. Fictional practice material is labelled. Dates record content updates; no independent expert review is implied. More about OffLadder.