Why I'm continuing to learn Data Analytics, when AI “Can Do It All”
Have you felt it? That almost-futile feeling when you realise AI can produce in seconds what would take you a week? When complex concepts aren't easily grasped by your busy, time-poor human mind, it's natural to question the value of learning.
Look, I'm going to be honest. Very early into my Business Intelligence and Data Analytics course from CFI, I had "a moment."
You may have a similar “moment” where you're staring at your screen, perplexed. In this scenario, I was wrestling with nested IF statements, wondering where the filter formula goes in the syntax, and thinking:
Why? Why am I doing this when ChatGPT could write this in under 30 seconds?
But I shake it off, and continue.
Crisis.
And then it happened again. I'm halfway through a module on DAX and Time Intelligence. I’m having to replay sections and pause on syntax creation, learning when to use what, and I'm thinking:
WHY? I can ask Claude or Copilot. Like, right now. I could describe what I want, paste in some data, and boom. It’s done. WHY am I spending evenings learning mechanics when AI can just, well… do it?
I genuinely questioned whether I'd made the right decision or if I was too late into the game. The AI hype is real, and it's easy to feel like you're learning a skill that's about to become obsolete. This led to an insurmountable lack of motivation to continue. Am I learning something obsolete?
Click.
During a data project at work recently, I was trying to use AI to help me clean and analyse data. I gave it my requirements, and it spat out a solution. Great! Right?
Except... I couldn't tell if it was actually good or accurate. I didn't know enough about it’s process to be able to evaluate whether the relationships made sense, if the normalisation was appropriate, if it had accurately interpreted the columns, or indeed if there was a simpler approach. (There was)
I had an answer, sure. But not the the right one.
AI is brilliant at providing solutions in lots of specific use cases - I can't count how many. But you need to know what questions to ask, what guidance to give, how to interrogate, and what is possible. You still need to make the data comprehensible and formatted. You need to understand what "good" looks like to engineer effective prompts. You need enough domain and contextual knowledge to spot when the AI is confidently wrong (because, I assure you, it will be).
"Learn the rules like a pro, so you can break them like an artist." — Picasso
The Real Value of Learning
It isn't about memorising syntax or chart types. It's about building the mental models that let me:
Ask better questions. Understand what is achievable to be able to prompt AI more effectively - you know what you're trying to accomplish.
Evaluate solutions holistically. AI might give you a technically correct answer to a question you didn't intend to ask. Sometimes it will hallucinate. You need to know enough to catch that.
Spot the gaps. Understanding the fundamentals helps you identify what's missing from an AI-generated solution - the context, the real-world messiness, and deep knowledge of your domain and company.
We should never blindly consume AI. Only if you understand the "why" behind the "what," is it a powerful companion.
Why I'm back at it
Sure, the course is teaching me practical skills in Excel, SQL, Python, Power BI, and Tableau. But what I'm also learning is:
How to think about data problems systematically.
Where AI would be helpful and where it will sink my time.
What questions to ask and what process to take when stakeholders say "can you test this hypothesis?"
How to identify whether a problem needs complex analysis or just a pivot table. (Seriously, guys, most problems can be solved with a pivot)
Closure
I am back into investing time completing BIDA because even if AI could do most of it, knowing what needs to be done is just as valuable as knowing how to do it. Maybe more. Will I feel the need to memorise every single syntax structure? Probably not. But I will memorise what can be done and know exactly how to retrieve a solution. And I'll build with AI as a partner to speed up the solutions. I’ll even post a few of my projects here.
AI can write the code. AI can give you a list of metrics it thinks is important for your business, from general context . But you still need to build and architect the solution, understand the trade-offs, and know when the "optimal" answer isn't actually the right one.
My advice (if you care to listen)? Don't stop learning just because AI can do the task. Don't just learn about how best to use AI tools. Learn so you can use AI better. The fundamentals aren't futile. They're your competitive advantage in an AI-augmented world… and you should still build, (faster), if that’s what you want to do.
