Dear younger me,
I'm writing this from my desk at Morningstar in Mumbai, where I spent the last three hours watching ChatGPT do something I used to charge clients ₹15,000 a day for. Not the entire job—but enough of it that I had to stop, step back, and think hard about what this means for you, for me, and for everyone reading this on the local train from Kalyan Central.
This isn't a panic letter. It's a reality check wrapped in honesty.
Six months ago, I was skeptical. I thought AI was hype. I thought my skills—data analysis, financial modeling, report writing—were safe because they required "human judgment." Today? I'm not so sure. And I think you need to hear this before you invest another three years becoming really good at something that a machine might do better by 2027.
So here's what I've learned. Not from think pieces. From actually sitting with ChatGPT, using it daily at work, and watching what it does to the way I work.
1. The "Speed Tax" Is Real—And It's Coming For Salaries
Let me start with the thing that keeps me up at night.
Last week, I spent four hours building a financial dashboard for a potential client. Pivot tables, VLOOKUP formulas, conditional formatting, the whole deal. The kind of work that used to be my bread and butter. This week, I asked ChatGPT to build the same thing—I described it in three sentences, and it gave me a Python script that does it in 40 seconds.
The quality? Almost identical.
Here's what worries me: if a task that takes me 4 hours can be done by a machine in 40 seconds, what happens to the price of that task? What happens to my salary?
The Speed Collapse
This is already happening. I know three data analysts in Mumbai who just took a 12–15% pay cut because their companies realized that the volume of work they were doing could be cut in half with AI assistance. Not because they got worse at their jobs. Because the jobs got faster.
And honestly? The companies weren't wrong. If ChatGPT can handle 30% of your work in 10% of the time, then either your work was underpriced, or your value just went down. There's no third option.
The speed tax doesn't just hit analysts. It hits copywriters, customer service reps, junior accountants, anyone whose job is mostly "produce output according to a pattern." And that's... a lot of jobs.
What This Means For Your Salary Growth
You know what salary growth used to look like? You get faster at your job, you produce more, you get paid more. Simple math.
Except now? If you get faster because of ChatGPT, your employer also sees the same speedup. So they don't need more people doing the same thing—they need fewer people doing more of it. The margin for negotiating higher pay because you're "more productive" just shrank.
I'm not saying it's impossible. But you can't compete on speed anymore. Full stop.
2. The "Judgment Tax" Is Where Real Work Still Lives
OK, so if speed doesn't matter, what does?
ChatGPT can write a financial report. It cannot decide whether your portfolio strategy is aligned with your life goals. It can generate 10 email templates to a difficult client. It cannot read the room and know which one will actually land.
This is where the real frontier is.
At Morningstar, I've started using ChatGPT to handle about 40% of my routine work—data cleaning, basic visualizations, email drafts, report templates. But the remaining 60%? That's judgment. That's asking the right questions. That's connecting a financial metric to a human problem.
And here's the pattern I've noticed: the people in my circle who are getting raises and moving up are not the ones who got faster at analysis. They're the ones who started asking better questions. The ones who could say to a client, "Your mutual fund returns look good on paper, but here's the tax inefficiency eating 2% of your gains every year." The ones who can translate numbers into decisions.
What Judgment Actually Means
Judgment isn't vague. It's not just "critical thinking" (which sounds nice on LinkedIn but means nothing). Real judgment is the ability to:
1. Spot what ChatGPT will miss. I asked it to analyze a fund's performance last week. It gave me a perfect technical report. But it missed that the fund's best-performing year was 2023, when small-cap stocks surged—and now everyone thinks the fund is great, when really the manager just rode a trend. A human eye caught that because we know what "trend-chasing" looks like.
2. Connect the dots across domains. ChatGPT knows finance. It knows psychology. But it doesn't know that your 35-year-old client with ₹50 lakh saved up will absolutely panic-sell in a market crash because they've already panicked once before in 2020. That judgment lives in experience.
3. Know when to break the rules. The standard asset allocation for someone your age is 70-30 stocks-bonds. But if you have a stable government job with guaranteed pension? Maybe 85-15 makes sense. ChatGPT will give you the standard answer. A good advisor breaks it intelligently.
The Real Value Prop In 2025
If you want job security and real pay growth, you need to stop being interchangeable. Right now, every data analyst in India can give you similar output. Make yourself one of the few who can explain what it means.
This is hard. It's uncomfortable. It requires you to develop opinions, test them against reality, and be wrong sometimes. ChatGPT can't do that because it doesn't have skin in the game. You do.
3. The Skills That Actually Matter Now (Spoiler: It's Not What You Think)
So if speed and pattern-matching don't matter, what does?
Let me show you what I'm actually investing time in right now—and what I'm not bothering with anymore.
| Skill I'm Learning Hard | Why | Skill I'm Letting Go |
|---|---|---|
| Asking better questions | ChatGPT can answer any question. It can't ask the right one. | Memorizing formulas |
| Domain depth (knowing ONE field really well) | AI is broad but shallow. You need to be deep in your area. | Broad skill collection |
| Communication (especially writing) | AI can't read the room. You can. | Presentation formatting |
| Building judgment frameworks | Decision-making is the last frontier. | Following best practices |
| Owning outcomes (not just tasks) | If something breaks, a human needs to explain why. | Task completion |
The "Domain Expert" Moat
Here's what I didn't expect: becoming more valuable to my employer didn't require learning more tools. It required knowing less—but knowing it way better.
I used to try to be good at everything—Excel, Python, SQL, dashboard tools, financial modeling, the works. Spread thin, not particularly exceptional at any one thing.
Now? I'm betting deep on one thing: understanding financial markets and investor behavior. And I'm letting ChatGPT handle the implementation. Want me to build a Python script? ChatGPT does it faster. Want me to explain why a market moved the way it did? That's where I'm valuable.
The moat isn't the tool. The moat is the understanding.
Communication In The Age Of AI
And honestly? Writing matters more now, not less.
When everyone can generate text, the people who can write specifically—for a specific person, with specific context, with actual conviction—stand out. ChatGPT writes perfectly generic emails. You can write the email that actually gets a response.
This is why I started this blog. Not because I'm a great writer (I'm not). But because I can write for you, specifically, in a way that ChatGPT can't—with real experience, real mistakes, real context from Kalyan and Mumbai and the actual financial world you're living in.
4. What You Should Actually Do (Starting Today)
OK, let's get practical. You're probably somewhere between "panicked" and "this doesn't affect me because I'm special." Both are wrong.
Here's what you should actually do:
Use ChatGPT Like It's A Co-Worker, Not A Threat
Start using it tomorrow. Seriously. Don't wait for the "perfect" moment. Your job right now is to figure out what parts of your work it can handle and what parts it can't. The people who do this early will adapt smoothly. The people who pretend it's not happening will get blindsided.
I spend about 2 hours a day with ChatGPT now. Not because I'm lazy, but because I'm trying to understand its boundaries. Where does it fail? Where is it surprisingly good? What can I outsource to it so I have more mental energy for the parts that actually require judgment?
Here's what I've learned: it's fantastic at anything with a clear output pattern. Terrible at anything requiring nuance or real creativity. Completely lost when it needs to say "I don't know" (it doesn't, it just makes things up).
Build A "Judgment Journal"
This is something I started doing three months ago, and it's quietly become the most valuable thing on my desk.
Every week, I write down 2–3 decisions I made that ChatGPT couldn't have made. Not big decisions—just professional judgment calls. "Client wanted to move to aggressive funds, but their behavior in 2020 showed they'd panic-sell. Suggested moderate instead." Or: "Dashboard template ChatGPT created looked good, but missed the tax inefficiency story. Rewrote the narrative."
I review these every month. It's a log of my actual value. And it's becoming increasingly clear which parts of my job are defensible and which ones are replaceable.
You should do the same. Track the moments where your judgment, your experience, your reading of the room—actually mattered. That's your future moat.
Invest In One Skill Deep, Not Five Shallow
Stop trying to be a generalist. Pick one domain. Go deep. Become the person in your company who really understands that thing. (For you, maybe it's financial analysis. Maybe it's the specific industry you work in. Pick something.)
Then, use ChatGPT to handle everything else.
This isn't new advice. But it's more true now than ever before. The people who survive the next decade aren't the ones who are "good at lots of things." They're the ones who are exceptional at something specific.
My Perspective
I've been analyzing financial data for seven years. Spreadsheets, models, reports—that's been my identity. Three months ago, I would have told you that AI couldn't do "real analysis." I was wrong. Confidently, publicly wrong.
What I got right was this: I saw early that the real value wasn't in the analysis itself. It was in the questions I asked before analyzing, and the decisions I recommended after. So instead of panicking, I started leaning into that. I'm using ChatGPT to handle the analysis faster, freeing me to think about the judgment layer.
At Morningstar, I noticed something surprising: the analysts who started using ChatGPT early are the same ones getting pulled into strategy meetings. Not because they're "more productive," but because their output became more valuable. They could give clients insights, not just dashboards.
That shift—from output to insight—is the actual job security move. And it only works if you use ChatGPT as a tool to amplify your judgment, not replace it.
Final Thoughts
Look, I know this is a heavy letter. You probably wanted me to tell you everything is fine and your job is safe. I can't tell you that because I don't know if it's true.
But here's what I do know: the people who panic are already behind. The people who ignore it are betting on luck. The people who lean in—who use ChatGPT, who invest in their judgment, who get strategic about their skills—those people are going to be fine. Better than fine.
You're young. You have time. Use it wisely. Stop being good at tasks. Start being great at judgment. The next five years in your career depend on it.
The train from Kalyan to Mumbai is about to get a lot more interesting.
—Dattatray
Dattatray Dagale
Data Analyst • Blogger • Mumbai
I'm a data analyst from Kalyan, Maharashtra, working at Morningstar. I write about personal finance, career growth, and everyday life for Indian millennials — the stuff I wish someone had told me earlier.
Written by Dattatray Dagale • 28 September 2026
.png)
.png)
0 Comments