Working with AI, Not Competing with It: A Guide for Analysts

With the continuation of AI tools evolving, one question keeps surfacing: 'Will AI replace analysts?'

From what I’ve seen through my own work — building data pipelines, deploying LLMs, and designing BI dashboards — the answer isn’t that simple. The real shift isn’t about replacement; it’s about reinvention.

Analysts aren’t being phased out. They’re being augmented — with tools that reduce the manual grunt work and free us up to focus on strategy, storytelling, and insight.

Lets dive into this article - a guide for anyone navigating this transition. Whether you’re starting out or looking to stay ahead, here’s what I believe today’s AI-ready analyst needs to thrive.

SQL and Python Still Power My Work — Every Single Day

When someone says, “AI is going to replace analysts”, I usually just smile. Because honestly? Most of my days still begin the old-school way — writing SQL queries, debugging Python scripts, and getting my hands dirty cleaning up messy data.

In my recent experience working as a Product Analyst at NoteSight, I recently built an end-to-end ETL pipeline to pull in user behavior data and load it into Snowflake. Nothing flashy — just practical automation that made our reporting faster and more reliable. In fact, it bumped our data reliability up by 10%.

The project was a strong reminder for me that, while AI is evolving fast, the fundamentals haven’t gone anywhere. SQL and Python are still how we ask questions and make sense of data. If you’re not fluent in those, no amount of AI prompts can truly help you uncover insight — it’ll just be noise.

Prompt Engineering: The Skill I didn't know I needed

One of the most rewarding experiences I had as a Graduate Assistant at Stevens was building a classification system for over ten years’ worth of faculty publications. Until then, the sorting process was painfully manual — hours spent labeling records and cleaning entries.

I decided to try something different. Using OpenAI’s LLM API and a bit of Python, I built a prompt-based pipeline that could automatically bucket the data. It cut the processing time down by over 70%, and more importantly, it worked consistently.

That project was an eye-opener. It made me realize that AI isn’t just a buzzword — it’s a tool, and how well it works depends entirely on how well you communicate with it. That’s where prompt engineering comes in. It’s not just about throwing a question at a model — it’s about knowing how to ask the right one. It's where logic meets language, and it's quickly becoming just as important as writing clean code.

Dashboards are turning into Conversations - Analyst Must Keep Up

I’ve built dashboards to track everything from credit usage to academic impact, and while visuals are still important, I’ve noticed something shifting: people don’t just want charts anymore — they want clarity.

At HighRadius, I remember presenting a Power BI dashboard that outlined client usage metrics. I was prepared to talk about numbers — but the room wanted something else. They weren’t asking, “What’s the usage rate?” — they were asking, “Why are users dropping off here?” and “What do we need to fix next?”

That’s when it clicked. The best dashboards don’t just report — they respond. And now with AI integration, we’re seeing tools that can analyze patterns, flag anomalies, and even suggest actions — all in real time.

As analysts, our role isn’t just to build dashboards anymore. It’s to design conversations — ones that guide decisions, not just display data.

Your Soft Skills Will Make or Break You

Let’s be real — tools like SQL, Python, Figma, or Power BI are powerful. But in every role I’ve taken on, they’ve only gotten me halfway. What’s made the real difference? The ability to work with people and translate complexity into clarity.

At NoteSight, I wasn’t just building wireframes in Figma — I was sitting down with designers, digging into user behavior, and figuring out why users were dropping off. At HighRadius, aligning onboarding workflows with consultants wasn’t about automation scripts alone — it was about understanding client pain points and communicating value clearly.

And that’s the part AI can’t do.

AI can analyze numbers. But it can’t walk into a room and explain to a stakeholder why something matters. It can’t pick up on hesitation in a client’s voice or tell when a product manager is worried about a launch. Context, empathy, and storytelling — that’s where your real value lives.

The best analysts I’ve worked with aren’t just technical. They’re thoughtful. Curious. Clear. That’s the kind of analyst I aim to be every day.

Being “AI-Ready” Means Staying Curious — Not Knowing Everything

Here’s the truth: I’m not an AI engineer. I don’t build models from scratch or write machine learning algorithms all day. But I’ve deployed LLMs to automate repetitive work, integrated Snowflake into my analytics pipelines, tested UI changes in Figma, and built dashboards that help real people make better decisions.

And through all of that, one thing has stood out: you don’t need to know everything — but you do need to stay curious.

Whether it’s spinning up an ETL workflow, experimenting with an A/B test, or writing a better prompt that helps an LLM give you something useful — the real power isn’t in mastering every tool. It’s in being open to exploring what works, iterating fast, and applying AI in ways that are actually meaningful to the business.

To me, being “AI-ready” isn’t about becoming some kind of tech wizard. It’s about asking, “How can I make this better?” — and not being afraid to try.

Closing Thoughts

AI is no longer some far-off concept — it’s already transforming how we analyze, interpret, and act on data every day.

The most impactful analysts I know aren’t the ones resisting change. They’re the ones leaning into it — using AI to automate the repetitive, uncover deeper insights, and focus on strategy over execution.

If you’re building a career in analytics, this is your moment to rethink your stack, reframe your role, and reimagine the kind of value you can create.

Thanks for reading!

💬 I’d love to hear your thoughts: What AI tools or workflows have you started integrating into your analytics practice? Let’s learn from each other.