The data scientist's evolution: Why AI is redefining, not replacing the role

If you're in the data space in any capacity, your social media feeds are probably filled with two completely opposing narratives: "The role of data scientists is going to disappear" vs "AI produces slop and is only good for entry-level work." The reality is somewhere in between. Data scientists aren't disappearing, but AI is definitely changing the role—and honestly, for the better.

Since being named the "sexiest job of the 21st century" over a decade ago, the number of data scientists has exploded. With that growth, supply has started to outpace demand, and if you spend any time on Reddit, you'll inevitably see posts asking if it's even worth getting into data science anymore.

At Fabi, we work with data teams at organizations of every size and across industries—from small bootstrapped startups to large public companies. And since we're building a product that puts an AI data analyst in the pocket of everyone in the enterprise, including data scientists themselves, we have a pretty unique vantage point on where this role is headed.

What is a data scientist, really?

If you're reading this, you might feel like this answer is obvious, but titles in the data space are so messy that I thought it was worth taking a minute to clearly define what I mean by data scientist as a role, not just a title.

When I say data scientist, I mean someone who has a strong technical foundation—especially in SQL and Python—good to great math and stats skills, and whose job is to help the business find those "aha" insights that increase product adoption and drive revenue growth. By this measure, there are plenty of folks on data teams who fit the bill even though their title might not be data scientist, and vice versa.

The uncomfortable truth: Most organizations don't actually need data scientists

Before we dive into how AI is changing the role, let's take a quick detour to talk about something important that nobody really wants to say out loud: Most companies don't need data scientists in the purest sense.

This reality has little to do with AI and everything to do with what businesses actually need. Most businesses are focused on the top and bottom line and are nowhere near the point where they need to be building regression models for sales forecasting. Instead, they need to be able to reliably report on sales numbers from the past quarter.

If you're the first data hire at a small or growing startup and your title is data scientist, this probably hits close to home. I'd venture to guess you're spending more time finding data, cleaning it, and doing what might traditionally be a data engineer's role, while spending way more time in spreadsheets and slides than you ever thought you would.

And that's totally fine! If your passion is going deep on custom models, then maybe this isn't the right role for you. But if you're excited about moving the business forward, this is what reality often looks like.

How AI is changing the role of data scientists

Let me lay down the facts that we know before we dig into the implications:

AI is becoming incredibly good at coding. A year or two ago, it might have just been a nice gimmick for less technical folks, but nowadays, with the right tool, it's 10Xing productivity for even the most technically advanced individuals.

AI doesn't understand the full context of your business. Maybe someday in the not-too-distant future, AI will connect to every business system, observe conversations, and get access to call recordings to gather context from human-to-human meetings. But until then, AI doesn't have access to all the subtleties of your business. If you get context about how "pipeline" is defined based on a discussion between your CMO and CRO in a meeting, unless AI has a way to tap into that discussion, that's context only you have.

What this means for your skills

If we consider these two facts together, it's pretty clear that coding skills are actually becoming less valuable. Just take Fabi as an example—extremely experienced data practitioners are using our AI analyst agent to generate anywhere between 90 to 95% of the code they run.

So rote memorization and knowing how to code in SQL or Python from scratch will be devalued. Instead, what's going to be valued is:

A deep understanding of the business. Literally knowing how the business makes money and generates growth or profit.

A deep understanding of stats and machine learning to supervise AI code. Someone needs to know whether what the AI generated actually makes sense statistically.

The ability to thread the needle from raw data to PowerPoint. This requires solid data modeling and data engineering skills—you need to understand the full stack.

The rise of citizen data scientists

Going back to the earlier point I made about most organizations not needing pure data scientists, we're seeing something interesting because of this shift: a rise in "citizen data scientists."

These are people whose full-time job may not be to do "data science," but instead are product managers, growth marketers, engineers, or even founders. With the help of AI, they're able to perform the right level of analysis they need to make decisions.

This isn't replacing data scientists—it's actually creating more demand in the long run.

Will the data scientist role disappear?

From here, the next logical question is: Does this mean the role of data scientists will disappear?

In my opinion: No, but most data scientists will need to adapt. Anyone hired as a full-time data scientist will need to be able to speak to the business fluently and also get involved in data modeling and data engineering.

Here's the thing though: I actually think the increased ease of access to data and insights across the organization is going to increase demand for this type of role in the longer term.

Let me give you an example. A head of product has a question about user activation. In the past, this data may have been just enough out of reach that this person used old data, no data, or simply their gut. (Unfortunately, this is way more common than it should be.) But now with AI, this person can get a directionally correct answer on their own in minutes if they have the data.

This helps them make better decisions, which drives adoption and growth. But it also helps them articulate the value of data and the need for data experts to make this data readily accessible and trustworthy. When they want to go deeper or need more sophisticated analysis, who do they turn to? The data scientist.

The time to adapt is now

If you're a data scientist (or thinking about hiring one), the good news is that AI certainly isn't taking your job tomorrow. But it is shifting where the value lives, and anyone in this role or contemplating it needs to think about how they adjust to the new reality.

And for hiring managers, you need to reconsider where the bottleneck truly is: Is it in writing SQL and Python, or is it in gathering data, understanding the business context, and modeling it correctly for the rest of the organization?

The data scientists who thrive in the next decade won't be the ones who can out-code AI. They'll be the ones who know what questions to ask, can validate the answers, understand the business deeply enough to provide real insight, and can communicate all of that effectively.

The role isn't disappearing—it's evolving into something more strategic, more business-focused, and honestly, more interesting. The question is just whether you're adapting with it.