Everyone keeps asking whether AI will replace analysts. I think that’s the wrong question entirely.
I believe AI has the potential to make analysts 10x to 100x more productive, not by replacing them, but by removing the drudgery that’s been standing between them and the work that actually matters: understanding the business, understanding the context behind the numbers, and getting the data right instead of wrestling with the mechanics of how to get there.
We’ve seen this movie before in software engineering. AI coding tools didn’t shrink the developer population. They made them dramatically more productive and shifted the job away from writing boilerplate code toward thinking more about the problem worth solving. And they handed those same powers to people who had never written a line of code in their lives.
Analytics is heading in the same direction.
The mechanical part of the job was never where the value lived. Knowing which button to click is not the same as knowing which problem to solve.
Curiosity is still the competitive advantage
I’ve always thought there are three kinds of questions.
The questions you already know how to answer. The questions you know you need to answer but haven’t yet. And the questions nobody thought to ask.
AI is incredibly good at helping with the first two. It can surface patterns faster than any human. It can summarize, calculate, compare, and generate ideas in seconds.
But the biggest opportunities in business usually aren’t hiding inside known questions. They’re hiding inside the unexpected.
That’s where human curiosity and a deep understanding of the business still matters most. That’s the part AI doesn’t replace.
General-purpose AI is not an analytics strategy
Many organizations are learning this distinction the hard way.
I saw this play out recently with a company that decided AI was going to solve everything. They rolled out a popular AI assistant to 300 analysts. It was ambitious, and I respect the intent. But by June, their AI budget was already gone. Worse, 300 people were asking the same questions over and over, burning through tokens, and getting inconsistent answers. That’s because data work needs determinism, and general-purpose AI wasn’t built for that.
Then everyone started creating their own artifacts and analyses with no way to share them. By the time anything was shared, it was already stale. What started as an AI-forward initiative turned into chaos.
I don’t share this story as an argument against AI. It’s an argument that a general-purpose assistant is not a purpose-built environment for scalable data analysis. That requires trusted data, business logic, repeatable methods, and deterministic results.
Trusted data starts with the analyst
There’s a related shift worth watching too. As AI becomes accessible to more people, business users will increasingly perform activities that once required specialist knowledge. They will ask questions in natural language, explore trends, and generate their own analyses.
That puts new responsibility on analysts as they increasingly become stewards of the trusted data, business context, and methods that allow others in the business to analyze with accuracy.
You don’t need perfect data before you begin. Waiting for perfection can become its own obstacle. But you do need data that is curated enough, understood well enough, and governed appropriately for the decision being made.
The better the foundation, the more useful the AI becomes.
Inventors will outperform imitators
Every major technology shift creates two kinds of companies.
The ones that recreate yesterday’s processes with today’s technology and the ones that invent entirely new ways of working.
We’ve seen it with cloud. We’ve seen it with digital transformation. We’re seeing it again with AI.
The winners won’t be the organizations with the biggest AI budgets. They’ll be the ones that rethink how decisions are made.
That’s how we’ve approached building Golden. We don’t believe AI replaces analysts. We believe analysts capture the business logic that makes AI trustworthy in the first place.
Because at the end of the day, AI doesn’t create understanding. People do.
So if you’re worried AI will replace analysts, I’d encourage you to consider a different possibility.
AI isn’t reducing the importance of the analyst. It’s making the best analysts the most valuable people in the room.