Over the past few years, I am sure we all have witnessed and experienced a slow but significant change in the way we work. We have upskilled ourselves often and my analytics stack kept rebuilding itself underneath me. I remember when I first read about the prowess of Generative AI, my thought stopped at “I will never have to spend hours on StackOverflow to debug my code again.”
At work today, I use AI models to write, design, code, sound board my ideas, complete tasks at breakneck speed, help me draft proposals for executives, create analytical stories, test my hypotheses, prepare me for important meetings and difficult discussions and so much more.
However, the more time I spend integrating AI to make my life “easy”, I realize that there is some tough work humans do that AI cannot exactly replicate. Generative AI cannot set aspirations, make decisions when times are tough, build trust among stakeholders, or hold itself accountable for good or bad.
That work remains deeply human.
As agentic AI absorbs more of the analytics workflow, the scarcity in your tech stack will not be speed or quality of execution but the ability to generate an original thought, apply business judgment, and decide what deserves attention in the first place.
Independent Thought Is Still a Competitive Advantage
I’ve come to believe that AI won’t replace human jobs. But AI can certainly help us execute work faster than ever before. At that point, the real question becomes: what do we do with the time that AI gives back to us? Do we use it to think more deeply, explore new ideas, and solve harder problems? Or do we gradually stop doing those things ourselves?
Compared to earlier this year, I am noticing a shift in my own behavior. I increasingly catch myself asking Copilot to interpret patterns in data before looking at it myself. I see myself preferring convenience over challenging myself. Why spend twenty minutes exploring a dataset when an AI can lay the ground for me in seconds?! The convenience is undeniable.
However, this dependency on AI is making me nervous. And that has nothing to do with the agent getting it wrong (because often it doesn’t). I get nervous thinking how easy it is to skip the most valuable part of analysis: forming my own initial point of view! I would have simply received someone else’s version of it, dressed up as mine.
That’s my big anxiety around the whole AI-and-jobs conversation: not me becoming obsolete, but unoriginal.
Imagine going through an entire career without regularly having a single original thought. I do not want the intellectual friction that produces insights, creativity and conviction to go away because I derive a great sense of accomplishment working through the challenges.
My thought is that in a world where intelligence becomes abundant, independent thought becomes scarce.
The Business Translation Problem
Working in analytics & AI for over seven years now, I know that the analysts who struggle are rarely the ones who aren’t the strongest in writing SQL queries or code in Python; it’s the analysts who can’t translate what the data says and what the business should do next.
Today’s AI can do far more than generate SQL or summarize dashboards. With access to semantic layers (a business-aware abstraction that sits with your organization’s AI model and helps AI understand metrics, definitions, relationships, and organizational context), modern agents can understand metrics, detect patterns, recommend actions, and in some cases, even execute actions autonomously (Agentic AI) with the knowledge of business rules, historical decisions and enterprise risks.
I am saying that the analytics stack is being rewritten because I am seeing AI become increasingly capable of both analysis and execution. And yet, a critical gap remains.
The challenge with using AI at work now isn’t understanding the business. Increasingly, AI can. The challenge is getting AI to own the consequences of business decisions (maybe we will reach there too tomorrow).
Humans routinely make decisions that contradict the data because they understand the strategic priorities, organizational politics, cultural nuances, regulatory risks, or simply a belief about where the business needs to go next. These decisions are not always objectively correct, but someone must ultimately be accountable for making them. But AI, in my opinion, cannot determine when the rules should be broken. That judgement is still entirely human.
So, the analytics stack may be changing but accountability isn’t and that’s the layer Agentic AI still can’t touch.
How I Use AI to Work Faster Without Letting It Think for Me
Where I’d Focus, Right Now and Later
As Agentic AI is taking over more of the work that used to consume our days from query writing to documentation, reporting, analysis, research and even framing recommendations into a story, the question is no longer whether AI can do parts of our job. The question is what we do with the time it gives back.
Do we reinvest that time into learning something new and deeper, develop a stronger judgment, or work on a more ambitious problem? Or do we gradually outsource those capabilities too?
For me, the real career question in this AI era is: what should I delegate to AI, and what should I deliberately keep my hands on?
This Quarter
-
Audit where you spend most of your time each week. Separate the work that is mechanical from the work that requires judgment.
-
Delegate the mechanical work to AI freely – write up long summaries, documentation, data preparation, first drafts, exploratory queries. Save your energy for questions that matter and what the answers to those questions actually mean.
-
-
Before asking AI for an interpretation, spend five minutes forming your own hypothesis.
-
Use the AI agents to challenge your thinking, not replace it.
-
Before The Year Ends
-
Learn how to collaborate with AI agents the way you would collaborate with a strong new hire or a junior analyst.
-
Context, constraints, goals, and feedback matter more than prompts.
-
-
Use AI as a coach rather than merely an executor. Ask it to pressure test your reasoning, expose blind spots, simulate stakeholder reactions, and explore alternative scenarios.
-
Practice translating complex findings into business decisions. The more AI commoditizes technical execution, the more valuable your communication and judgment becomes.
Over the Next Few Months
-
Convert your experience into frameworks, principles, and decision-making models that can scale beyond you.
-
Seek opportunities at the intersection of business strategy, analytics, and AI and keep reading about all the advancements in this landscape
-
Continuously invest in the skills that are very human and get strengthened through use (but weakened by delegation) like critical thinking, creativity, judgment, leadership, and vision.
Final thoughts: As AI takes over more execution, the human advantage shifts to judgment, aspiration, and independent thought.
···
That’s it from my end on this blog post. Thank you for reading! I hope you found it an interesting read!
Rashi is a data wiz from Chicago who loves to analyze data and create data stories to communicate insights. She’s a full-time senior healthcare analytics consultant and likes to write blogs about data on weekends with a cup of coffee.
···