You learned SQL because data felt like the safe bet. Now a chatbot answers your manager's questions. What is happening to the work and to you.
This in-depth guide covers everything you need to know about will ai replace data analysts? what the hiring data shows. Based on verified income data and real-world case studies from our database of 138 side hustle tactics.
AI is already doing a large share of routine data analyst work: writing the SQL, pulling the numbers and building the first draft of the dashboard. Data analysts and data scientists are still being hired, and the official forecasts still show strong growth for data scientists, but the entry-level door has narrowed sharply. As of October 2026, Indeed's data and analytics job postings sit lower, measured against their own pre-pandemic level, than any other field Indeed tracks.
The course you finished at midnight
You learned SQL at the kitchen table after the kids were asleep. Maybe you paid for a bootcamp with savings that were meant for something else, or you did the online certificate in the evenings while working a job you did not like. Everyone said data was the safe bet. Every company has data, every company needs someone to make sense of it, and the pay was good enough to move out, pay down the loan, maybe send something home to your parents.
You built a portfolio. A Tableau dashboard on housing prices, a Python notebook on churn. You sent out applications. Now the replies are slow, the "junior" roles ask for three years of experience, and your manager at your current job has started typing questions into a chatbot that answers in a chart. You wonder whether you trained for a job just as software learned to do it.
What AI can already do with your data
The pitch from every large data platform in 2025 and 2026 has been the same: let anyone in the company ask a question in plain English and get the answer without waiting for an analyst.
Snowflake. In November 2025, Snowflake made its Snowflake Intelligence agent generally available and said more than 1,000 customers had used it to deploy over 15,000 AI agents in the previous three months. The product is built so any employee can ask a business question and get an answer drawn from the company's databases. Snowflake's own engineers reported in August 2024 that its Cortex Analyst text-to-SQL system reached "90%+ SQL accuracy on real-world use cases" on an internal test set, against 51% for a single prompt to GPT-4o. That is the company's own test, so treat it as a claim. It still shows where the effort is going.
Databricks. When Databricks made its AI/BI Genie tool generally available in June 2025, it said over 4,000 customers adopted it during the preview. The customer quote it chose to publish is the clearest statement of the problem for analysts. Shahmeer Mirza, senior director of data, AI/ML and R&D at 7-Eleven, described users "constantly pinging analysts" with questions such as how sales in the southwest region compared with the previous year. Then he said: "The idea of being able to just ask Genie, rather than hunt for the right analyst and hope they get the answer right, has been very exciting for the business."
Read that again from your side of the desk. The "pinging" he describes is a large part of what a junior analyst does in their first two years. It is how you learn the business, and it is how you prove you are useful.
Notice what these tools are aimed at. The person who designs the data warehouse or argues with finance about how revenue is counted is safe for now. The target is the request queue: the ad hoc question, the weekly export, the "can you just pull" message on a Friday afternoon. For a lot of junior analysts that queue is the job. It is also the training ground where you learn which tables lie, which teams double count and which numbers the chief executive actually reads. When the queue goes to a chatbot, the training ground goes with it.
Julius AI. At the small-business and individual end, Julius AI said in July 2025 that it had over 2 million users who had created more than 10 million data visualizations. Its chief executive, Rahul Sonwalkar, said: "Much like Excel became essential in the 90s, Julius is becoming the standard for today's business leaders." A founder who once hired a freelancer to clean a spreadsheet and chart it now uploads the file and asks.
The people who do this work have already changed how they work. In dbt Labs' 2025 survey of 459 data practitioners and leaders, 80% said they use AI in their daily workflow, up from 30% the year before, and 70% of those daily users use it to write code.
A Microsoft Research team that studied how people actually use Bing Copilot at work scored occupations by how much of their work AI can apply to. In the July 2025 paper's table of the 40 most exposed occupations, data scientists score 0.357 and market research analysts and marketing specialists score 0.350, in the same table as mathematicians (0.386) and news analysts, reporters and journalists (0.383). If you read our post on mathematicians, you have seen how fast the top end of quantitative work is moving.
Question to sit with: if your manager could get 80% of your weekly reports by typing a question, what is the other 20% that only you do, and does anyone above you know it exists?
The numbers on hiring
Here is the part that hurts. Indeed publishes a daily index of job postings by field, with February 1, 2020 set at 100. On October 2, 2026 the data and analytics index stood at 63.1, the lowest of all 47 occupational sectors in Indeed's data. Postings across all US jobs were at 99.4 on the same day. The data and analytics line peaked at 202.7 in March 2022, during the hiring boom, so postings have fallen about 69% from that peak.
The fall did not stop in 2025. Indeed's data showed data and analytics job posts down 13% year over year through October 2025. Laura Ullrich, director of economic research at Indeed's Hiring Lab, told CIO Dive in November 2025: "AI didn't cause the bust in hiring in the tech sector, but it might be preventing it from recovering at the same rate it would have."
That sentence is the fairest summary we found. The boom of 2021 and 2022 over-hired, and the correction would have happened anyway. What AI changes is the recovery. Companies that would normally hire their way back up are now asking whether a tool can do the work first.
The jobs that remain look different. In January 2026 Hiring Lab reported that 45% of data and analytics postings mentioned AI as of December 2025, the highest share of any sector it analysed. Tech postings that mentioned AI were about 45% above their February 2020 level, while total tech postings were 34% below. The work is moving toward people who build and supervise the AI tools, and away from people who answer one-off questions.
Who is losing the first rung
The clearest research on entry-level jobs comes from Stanford's Digital Economy Lab. Erik Brynjolfsson, Bharat Chandar and Ruyu Chen used payroll records from ADP. In their revised paper of August 2026 they report that employment of young workers (ages 22 to 25) in AI-exposed occupations now stands 19% below where it would be had it kept pace with less exposed jobs, and that "experienced workers show no comparable gap." The authors call this a descriptive finding, and the gap is measured against a comparison group. It still means the people losing out are the ones at the start, which is where most data analysts are.
Bootcamps were the other main door into data work, and the largest operator walked away. In December 2024, 2U, which ran university-branded bootcamps that produced some 96,000 graduates, announced it would leave the bootcamp business. Interim chief executive Matt Norden said: "Demand for entry-level tech roles has decreased while the pool of available, experienced tech talent has expanded." He added that "the long-form, intensive training that boot camps provide no longer aligns with what the market wants and needs."
We looked for audited job placement rates for data analytics bootcamp graduates in 2025 and 2026 and could not find any we trust. Many bootcamps publish their own figures, and the methods vary from school to school. If a bootcamp is quoting you a placement rate, ask how many graduates were left out of the calculation and what counts as an "in-field" job.
Question to sit with: when you paid for the course, were you buying a skill, or a ticket to a hiring market that has since shrunk? The answer changes what you do next.
What the official forecasts say
The forecasts are kinder than the job boards, and it matters to understand why.
| Figure | Value | Source |
|---|
| Indeed data and analytics postings index, Oct 2, 2026 (Feb 2020 = 100) | 63.1, lowest of 47 sectors | Indeed Hiring Lab data |
| Same index at its March 2022 peak | 202.7 | Indeed Hiring Lab data |
| Change in data and analytics posts, year to Oct 2025 | down 13% | CIO Dive, Nov 2025 |
| Data and analytics postings mentioning AI, Dec 2025 | 45% | Indeed Hiring Lab, Jan 2026 |
| Young workers (22 to 25) in AI-exposed jobs vs comparison | 19% below | Stanford Digital Economy Lab, Aug 2026 |
| Data scientists, projected growth 2025 to 2035 | 35% (275,600 jobs in 2025) | BLS |
| Operations research analysts, projected growth 2025 to 2035 | 12% (113,100 jobs) | BLS |
| Market research analysts, projected growth 2025 to 2035 | 7% (952,700 jobs) | BLS |
| All US occupations, projected growth 2025 to 2035 | 3.5% | BLS, Aug 2026 |
| Data practitioners using AI daily | 80% (up from 30%) | dbt Labs, Apr 2025 |
| Julius AI users | over 2 million | Pulse 2.0, Jul 2025 |
The US Bureau of Labor Statistics released its 2025 to 2035 projections in August 2026. It expects data scientist jobs to grow 35%, with about 24,800 openings a year and a median wage of $120,230 in May 2025. Operations research analysts are projected to grow 12%, with about 7,500 openings a year and a median of $88,940. Market research analysts are projected to grow 7%, with about 82,000 openings a year and a median of $78,760. Across the whole economy, BLS projects growth of just 3.5%.
The World Economic Forum's Future of Jobs Report 2025, based on a survey of employers, also puts "Data Analysts and Scientists" among the 15 fastest growing jobs to 2030, alongside big data specialists and AI and machine learning specialists.
So why does it feel so different on the ground? There are three reasons.
First, "data analyst" is not a single BLS occupation. Analysts are spread across many job codes, from management analysts to financial analysts to market research. The 35% headline belongs to data scientists, about 275,600 jobs in 2025, and those roles tend to ask for more statistics and programming than a typical analyst job.
Second, projections describe the end of a decade. They say little about the next two years, or about how many people are competing for each opening. A field can grow while far more graduates enter it than there are junior roles.
Third, forecasts count jobs and say nothing about who gets them. If companies hire the same number of data scientists but stop hiring the juniors who used to grow into them, the BLS line still goes up. The Stanford result points in exactly that direction.
Question to sit with: are you aiming at the job the forecasts describe, or the job the forecasts quietly assume someone else trained you for?
Where the evidence cuts the other way
There is honest counter-evidence. In the same dbt Labs survey, 40% of data teams reported growing in size, up from 14% the year before, and AI tooling was the top area for new investment. Companies that put AI agents on top of their data warehouses need people to define metrics, clean tables and check what the agent says, because a confident wrong answer to "what were sales last quarter" can cost real money. Ullrich also said in November 2025: "We still need a lot of people that have technical skills, and those people are needed across all the sectors and the economy."
The problem is that this demand sits mostly with people who already have experience. It does not do much for the person trying to get the first job.
What this means for you this month
If you are already working as an analyst, your safest move is to become the person who owns the definitions, the data quality and the judgment calls, because those are the parts the tools still lean on a human for. Keep a short record of decisions you made that changed what the business did. That record is what protects you in a restructure.
If you are still trying to break in, be honest about the numbers. The junior market is the weakest it has been since at least 2020, and a second bootcamp is unlikely to fix that. Keep applying, but build income alongside it. The skills you already have sell well in small chunks: cleaning and reporting for small businesses, or teaching SQL and Excel through online tutoring. People who can test and grade AI output on data tasks are also being paid for it; our guide to AI training jobs explains how that work is found and what it pays. If you are thinking about a bigger move, read career change at 40 before spending more on retraining.
For the wider picture, see jobs AI will replace, the state of the 2026 job market, and our post on whether AI will replace software engineers, where many of the same companies are making the same choices.