The Data Analyst Skills That Actually Matter Now That AI Can Write Your SQL

A few years ago, learning to be a data analyst meant learning to type. You memorized SQL syntax, you learned Excel formulas by heart, you practiced Python loops until they felt natural. Today you can open ChatGPT, Claude, or Gemini, describe what you want in plain English, and get a working query back in seconds.

So a real question people are typing into Google right now is simple: do I still need to learn SQL, Excel, or Python at all, if AI can write it for me?

Think of it like driving with GPS

A GPS app is great. It picks the route, it warns you about traffic, it even talks you through turns. But if you have never learned to drive, and you have no idea how roads and directions work, a GPS will not save you. The moment it glitches, sends you down a closed road, or points you the wrong way on a one-way street, you are stuck. You need enough driving knowledge to notice something is wrong and correct it yourself.

AI tools for data analysis work the same way. They are an excellent GPS. But you still need to know how to drive.

What is actually changing

AI can now write a working SQL query from a plain English request like "show me revenue by region last quarter." It can build an Excel formula, summarize a dataset, or draft a Python script for cleaning data. That part of the job, the typing and syntax, is genuinely getting automated.

This is showing up in the job market. Several 2026 hiring reports point to a real drop in entry-level analyst postings, roughly a third fewer than a couple of years ago, concentrated in junior roles whose main job was building the same recurring report or dashboard every week. That is the part of the work AI does well, so companies need fewer people doing only that.

The honest nuance

Here is the catch. AI-written SQL or formulas are only useful if you can actually read them and catch what is wrong. There is a well known example making the rounds of a junior analyst presenting AI-generated numbers to executives, only to discover mid-meeting that the query had a subtle bug in how it filtered dates. Six months of trend analysis, wrong, because nobody who understood SQL well enough checked the query before it went into the slide.

AI is a force multiplier. It multiplies whatever knowledge you already bring to it. If you understand SQL, AI helps you work faster. If you do not, AI just helps you make mistakes faster and with more confidence.

What this means if you are learning right now

Keep learning SQL, Excel, and Python fundamentals. Not because you will type every query by hand forever, but because you need enough of the underlying logic to review AI output and trust your own judgment over a machine's confident guess. Learning the fundamentals first, then using AI as a shortcut, is very different from using AI as a crutch from day one.

Once you have that base, spend more of your time on the things AI genuinely struggles with: understanding your company's specific business context, figuring out which question is actually worth asking, and explaining what a number means to someone who does not work with data every day. Those are the skills that are becoming more valuable, not less, as the typing gets automated.

If you are not sure where your own skills stand against what today's data and AI roles actually need, Flexing Data's AI Readiness Assessment is a low-key way to find out, and our Labs are there if you want hands-on practice with real datasets instead of just theory.