Being a Data Analyst in the Era of AI

In this Fabric Data Days session, I joined Eugene Meidinger to explore one of the hottest topics in our field: how AI is changing what it means to be a data analyst. We dove deep into practical strategies, real-world examples, and frameworks you can use starting today.

The AI Landscape for Analysts

Let’s be honest—AI is everywhere right now, and it can feel overwhelming. But here’s the thing: as data analysts, we’re actually in a prime position to benefit from these tools.

Why AI Matters Now

The pace of AI evolution is staggering. What was impossible two years ago is now table stakes. LLM capabilities are improving rapidly, and the tools we dismissed early on deserve a second look.

Key Insight: Re-evaluate AI tools regularly. What didn’t work six months ago might be exactly what you need today.

Generative AI for Analysts

We covered the rise of generative AI and where it fits in the analyst toolkit:

Natural language as a programming layer — You can now describe what you want in plain English
Automation opportunities — Repetitive tasks that used to eat hours
Documentation and cleanup — AI excels at tedious but important work

Practical AI Applications

This wasn’t a theoretical discussion—we focused on real workflows.

Requirements Gathering and Mockups

One of my favorite uses: having AI help create mockups and wireframes during requirements gathering. Describe what stakeholders want, and AI can generate visual starting points.

AI in Power BI and Fabric

We explored the AI features already built into your tools:

Model cleanup and documentation
Iteration on designs using style guides
Bridging the gap from problem statement to solution

Code Translation and Learning

Eugene demonstrated converting M code to PySpark—something that would take hours manually. AI handles these translations well, especially for:

Converting between query languages (M, SQL, PySpark)
Explaining unfamiliar code
Learning new languages through practical examples

The “Diamonds in the Dumpster” Framework

We introduced a practical framework for deciding when AI helps and when it hurts.

The Four Quadrants

Think of tasks along two axes:

High value vs. Low value — How much does this task matter?
High effort vs. Low effort — How hard is it without AI?

Quadrant 1: High Value, High Effort (Diamonds) These are your strategic projects. AI can help, but human judgment is critical.

Quadrant 2: Low Value, High Effort (The Dumpster) Perfect for AI. Let it handle the grunt work—code translation, documentation, boilerplate.

Quadrant 3: High Value, Low Effort Quick wins you should still do yourself. Don’t overcomplicate with AI.

Quadrant 4: Low Value, Low Effort Either automate completely or question if it needs doing at all.

Mapping Your Tasks

Look at your weekly work through this lens. Where are you spending time that AI could reclaim? Where does human expertise remain essential?

Why Analysts Still Matter

Despite all the AI hype, data analysts aren’t going anywhere. Here’s why:

AI is Like a New Employee

Treat AI as a new team member who’s incredibly fast but needs guidance:

It doesn’t know your business context
It can be confidently wrong
It needs supervision and verification

The Human Edge

What analysts bring that AI can’t replace:

Business context and institutional knowledge
Stakeholder relationships
Judgment calls on ambiguous situations
Knowing which questions to ask

Tools We’re Using

We shared the actual tools in our workflows:

LLM Agents and MCP — For model organization and complex tasks
Azure OpenAI and Foundry — Enterprise-ready AI deployment
OpenRouter — For comparing different models
LM Arena — Testing models head-to-head

Choosing the Right Model

Not all LLMs are equal for every task:

Frontier models (GPT-4, Claude) — Best for complex reasoning
Smaller models — Faster and cheaper for routine tasks
Specialized models — Fine-tuned for specific domains

Key Takeaways

Re-evaluate AI tools regularly — The landscape changes fast
Use the quadrant framework — Focus AI on high-effort, lower-value tasks
Treat AI as a new employee — Supervise, verify, and provide context
Start small — Pick one workflow and experiment
Keep learning — Your role is evolving, not disappearing

What’s Coming Next

AI capabilities will only accelerate. The analysts who thrive will be those who:

Build AI into their daily habits
Maintain their human judgment and business context
Stay curious and keep experimenting

Resources

Want to level up your Fabric and Power BI skills? Check out Training.tips for comprehensive courses.