Two weeks ago we talked about Nano Banana—Google's new AI image model that can edit photos with scary accuracy. This week, OpenAI released Sora 2 (video generation model) as well as AgentKit (and announced much more during OpenAI DevDay). Every day there's a new AI tool that can "do your job in seconds."

Let’s be honest, even with 12 years of experience in the industry - I get overwhelmed and anxious. Should I be learning n8n, LangChain, AI engineering? Would my role become obsolete in the following years?

So let's address the elephant in the room: Should data analysts be freaking out right now?

Short answer: No. But there's a longer, more nuanced answer you need to hear.

The Panic Is Real (And Totally Understandable)

I get it. You open LinkedIn and see:

• "AI can now write SQL queries for you!"

• "ChatGPT just analysed my entire dataset in 30 seconds!"

• "Why hire a junior analyst when AI can do it for free?"

It's enough to make anyone nervous. Especially if you're just starting out or switching careers.

But here's what the panic merchants aren't telling you.

What AI Actually Replaces

Let's be brutally honest about what AI is genuinely good at:

Data cleaning and preparation. The tedious "janitorial work" of analytics—the stuff we all hate anyway. AI can find missing values, standardise formats and prep datasets faster than any human. However, from personal experience, you still need to check and double check what exactly AI is removing, as once it cleaned up all my outliers which were VIP sales 🙄.

Basic reporting. Simple dashboards, routine reports, standard visualisations? AI can churn these out while you sleep.

Initial exploratory analysis. AI can quickly identify patterns, outliers and basic correlations in your data.

Writing simple code. Need a standard Python script or SQL query? AI's got you covered. I also use it to optimise query that I wrote, or explain to me query written by someone else. So many hours saved!

Here's the thing: if your entire job consists of ONLY these tasks, then yes, you should be concerned. But that's not what being a data analyst actually is.

What AI Can't Replace (And This Is Where You Win)

Data analysts are storytellers, translating complex data into actionable insights that decision makers can understand. And AI is terrible at this.

Business context. AI might identify a correlation, but it takes a human analyst to understand whether it's meaningful and relevant to the business. Is this trend seasonal? A market anomaly? A fundamental shift in consumer behaviour? AI doesn't know.

When I worked in finance, I learned that you can't just remove outliers because they "mess up your chart." In accounting, every number tells a story—sometimes the "weird" data point is the most important one. AI doesn't understand this nuance.

Asking the right questions. AI answers the questions you ask. But knowing WHAT to ask? That's pure human intuition combined with business understanding. For example, you can ask AI to overlay seasonality data with holiday calendar, or check some special events (like elections) that were happening at that time.

Stakeholder management. Remember my police project story? The sergeant didn't use Excel. No amount of AI will teach you how to read the room and adjust your communication for different audiences.

Ethical considerations. AI systems can inadvertently perpetuate biases present in the data they're trained on. Someone needs to catch this. That someone is you.

Just to remind you - AI is built on mathematical probabilities. In other words, if it was trained on a skewed dataset, it might be biased simply because it doesn’t know different. That’s why it is important to fact check AI results. The other day I saw a LinkedIn article with book recommendations, and some of the books don’t exist! AI just made them up.

Strategic thinking. You understand how changes in one area (like cost of sales) impact everything else (expenses, inventory, income). AI sees correlations. You see cause and effect.

The Real Picture: Job Growth, Not Job Loss

Here's what the data actually shows:

The US Bureau of Labor Statistics estimates data analyst positions may surge 36% between 2023 and 2033—significantly higher than the average job growth.

According to Alteryx's 2025 report, 70% of analysts say AI automation enhances their work effectiveness, while 87% feel more strategically valuable than ever before.

PwC's 2025 Global AI Jobs Barometer found that wages are rising for AI-powered workers even in highly automatable roles, suggesting AI is making workers MORE valuable, not less.

Translation: AI isn't replacing analysts. It's elevating the role.

Skills That Become MORE Valuable With AI

This is the game-changer. As AI automates the mundane, certain human skills become exponentially more valuable:

Communication skills. With AI handling the boring work, your ability to explain insights becomes your superpower. Presentation skills now appear in 14% of job postings - up significantly from previous years. By the way, grab my free Data Storytelling Guide if you haven’t done already.

Domain expertise. 69.3% of postings now seek domain experts with specific industry knowledge. Knowing healthcare, finance, retail or whatever your field is? That's your moat.

Critical thinking. AI gives you answers. You need to know if those answers make sense. Can you spot when AI is confidently wrong?

Creativity in problem-solving. AI follows patterns. You can think outside the box and approach problems from angles AI would never consider.

Curiosity. The analysts who thrive ask "why?" five times. AI stops at the first answer.

How to Position Yourself in an AI-Enhanced World

Stop fighting AI. Start leveraging it.

1. Learn to use AI tools. Machine learning mentions in job postings have doubled from 7% to 14% in 2025. You don't need to be an AI expert, but you should be comfortable using AI tools. Think of AI as your junior analyst—great at grunt work, needs supervision.

2. Double down on human skills. Communication, storytelling, business acumen, stakeholder management. These are your competitive advantages.

3. Become a domain expert. Pick an industry and go deep. The more you understand the business context, the more valuable you become.

4. Focus on the "why," not just the "what." Anyone can tell you sales dropped 15%. You need to explain WHY and what to do about it.

5. Build in public. (Remember last week's advice?) Show you're adapting, learning, and growing with the technology.

The Bottom Line

AI is automating the mundane, but it can't automate human judgment, curiosity, or business acumen.

The future isn't "AI OR humans." It's "AI AND humans working together."

The analysts who'll struggle are those doing rote, repetitive work without adding strategic value. The analysts who'll thrive are those who use AI to handle the boring stuff so they can focus on the interesting problems—the ones that require judgment, creativity, and deep business understanding.

Think of it this way: calculators didn't eliminate accountants. Excel didn't eliminate financial analysts. AI won't eliminate data analysts. It'll just change what we spend our time on.

And honestly? That's a good thing. I'd much rather spend my time solving interesting business problems than cleaning messy CSVs. Also, some of those CSVs and databases …. good luck, AI with cleaning that up!

Keep pushing 💪

Karina

I am working on something exciting….

A 14-day Data Portfolio Challenge where you'll analyse realistic Facebook ad data (not another Titanic dataset 🙄).

Here's what makes this different:

• Realistic, messy marketing data that every business uses

• Daily step-by-step guidance (no more "figure it out yourself")

• Use ANY tool you prefer - Excel, Python, SQL, Power BI

• Applicable to any industry (finance, e-commerce, healthcare)

• Walk away with a complete portfolio project

Early bird pricing: $99 (regular price will be $129)

Join the waitlist - link

Python tip

Cap Outliers Instead of Deleting them

What it does:

Finds 1st and 99th percentiles from YOUR data

Caps extreme values to these thresholds

Keeps 98% of data completely unchanged

• Preserves sample size (no deleted rows!)

• Data-driven thresholds (not arbitrary numbers)

• Handles both high AND low outliers

• Perfect for charts that aren't dominated by extremes

• Models train better on cleaned continuous data 

Perfect for:

• Financial data with crazy transactions

• Survey responses with impossible values

• Sensor data with measurement errors

df['score'] = df['score'].clip(
    lower=df['score'].quantile(0.01), 
    upper=df['score'].quantile(0.99)
)

# Conservative: 5th-95th percentiles
df['score'].clip(lower=df['score'].quantile(0.05), upper=df['score'].quantile(0.95))

# Aggressive: 0.5th-99.5th percentiles  
df['score'].clip(lower=df['score'].quantile(0.005), upper=df['score'].quantile(0.995))

Grab your freebies if you haven’t done already:

Data Storytelling Guide

Data Playbook (CV template, Books on Data Analytics and Data Science, Examples of portfolio projects)

Need more help?

Just starting with Python? Wondering if programming is for you?

Master key data analysis tasks like cleaning, filtering, pivot and grouping data using Pandas, and learn how to present your insights visually with Matplotlib with ‘Data Analysis with Python’ masterclass.

Grab your Pandas CheatSheet here. Everything you need to know about Pandas - from file operations to visualisations in one place.

Get CV, Portfolio and LinkedIn Review.

More from me: YouTube | TikTok | Instagram | Threads | LinkedIn