Dear younger me (or hey, friend in Kalyan),
Six months ago, I was drowning. Not dramatically — more like that slow-motion suffocation where you're busy but nothing feels meaningful. I was spending 8 hours a day in Excel, Slack, and email. Real work? Maybe 2 hours of actual thinking, analysis, decision-making.
Then I started experimenting with AI tools. Not the flashy stuff. Not ChatGPT for writing motivational quotes. Real, boring, daily-grind tools that actually moved the needle.
By month two, I had my evenings back. I was finishing work by 6 PM instead of 8 PM. And the quality of my analysis? Better. Sharper. Less noise, more signal.
This isn't a "10x overnight" story. That's Instagram nonsense. This is about being honest about what works, what's overhyped, and what's actually worth the 15 minutes it takes to set up. If you're between 22 and 35, living in Maharashtra, and tired of the commute eating your life, this matters.
The Real Problem Nobody Talks About
Let me start with what most AI productivity posts get wrong. They show you someone writing a novel in 2 hours or building a startup in a weekend. Cool. Irrelevant.
Your actual problem is different. You're not lacking ideas. You're drowning in *busy work*. The Slack messages at 4:47 PM. The email that requires three back-and-forths instead of one clear answer. The 45-minute meeting that could've been a 5-minute voice note. The spreadsheet analysis you've done 50 times already but management wants a slightly different angle this month.
That's where AI actually helps. Not by replacing you. By eliminating the nonsense between you and the real work.
When I started at Morningstar, I thought productivity was about doing more. Running faster. Taking fewer breaks. Stupid, honestly. What I learned: productivity is about reducing friction between you and deep work. Every minute spent on format, copy-paste, email threading, or data reorganization is a minute stolen from actual thinking.
AI is brutal at replacing thinking. It's amazing at removing friction.
The Commute Problem
Here's something specific to us Kalyan folks: you're spending 2 hours every day on that train. That's 10 hours a week. 40 hours a month. Roughly 480 hours a year. If you're making ₹10 LPA, that's ₹2,30,000 in lost productivity time. Per year.
Now, you can't do deep work on the local train. But you *can* use those 2 hours to handle the friction tasks — voice notes to transcribe, documents to review, emails to draft. If AI can compress what takes 45 minutes into 15 minutes, you just bought yourself 30 minutes of actual time back daily.
The Mental Load Is the Real Tax
Nobody counts this. Context switching costs something like 23 minutes per switch (studies vary, but the number doesn't matter — what matters is it's substantial). You're not just losing time switching between Slack, email, and your actual work. You're losing the mental state that lets you think clearly.
Every context switch is a small tax on your brain. AI can reduce switches. That's where 10x comes from.
The Three AI Tools That Actually Changed My Workflow
I'm not going to list 47 tools. That's not helpful. Here are the three that moved my needle enough that I'd use them even if they weren't free.
1. ChatGPT (or Claude) for Email and Slack Drafting
This is the boring one. Unsexy. But it saves roughly 45 minutes a week.
Here's what I do: when I need to write an email that requires more than 2 sentences, I open ChatGPT and write a rough outline. Something like: "Boss asked for fund performance analysis for Q3. Need to explain why large-cap underperformed, highlight our contrarian bets that worked, suggest rebalancing. Tone: professional but not robotic."
Takes 60 seconds. ChatGPT returns a draft in 20 seconds. I spend 3 minutes editing it for my voice and specifics. Done. Previously? I'd stare at the email for 8 minutes, rewrite it twice, sit with it for 10 minutes wondering if it sounds right.
The math: 5 such emails per week. 6 minutes saved per email. That's 30 minutes. Across a month? 2 hours. Across a year? 100+ hours. At ₹40 LPA, that's roughly ₹1,920 in reclaimed time per year. Not earth-shattering individually, but it compounds.
And honestly? My emails are clearer now. Because I'm not writing tired at 5:30 PM. I'm editing fresh.
2. Perplexity for Research
At Morningstar, research is constant. "What's happening with pharma valuations?" "How are IT services companies positioning for AI adoption?" "What's the regulatory outlook on crypto in India?"
I used to Google, click 6 articles, read 20 minutes worth of content, synthesize. That was my workflow in 2023.
Now I use Perplexity. I ask: "Why are pharma stocks trading at historically low valuations despite strong earnings growth in India?" I get a sourced answer within 20 seconds. It cites sources. I can verify.
Time saved: roughly 15 minutes per search. I do 4-5 of these per week. That's 60-75 minutes per week of research overhead gone.
The gap: Perplexity isn't perfect. It sometimes hallucinates sources. I still need to verify critical facts. But for the first-pass understanding? For "what am I missing here?" questions? It's phenomenal.
3. Claude for Data Analysis (the Big One)
This is where it gets real. And possibly controversial.
I work with datasets. Quarterly returns, fund performance, market data, all in CSVs or Excel. The boring part: cleaning, transforming, creating the right visualizations. This used to take 60-90 minutes per project.
I uploaded a sample dataset to Claude. Asked it to clean the data, remove outliers, create a pivot table showing returns by fund category and time period. Provide R-squared for correlation analysis.
8 minutes later, I had Python code. I ran it. Worked. Did the same thing I would've spent an hour building in Excel in 2 minutes.
Now, I'm not saying I don't understand the code. I read it. Verify it. Modify it when needed. But the scaffolding? The boilerplate? Gone.
This single tool might add 4-5 hours to my week of actual thinking instead of mechanical work.
How I Actually Integrated These (No Hype, Just Reality)
Week 1: The Honeymoon
I used Claude for everything. Drafting analysis notes. Creating presentation decks. Writing strategy documents. I thought I'd found cheat codes.
By day 4, I noticed: the output was good, but it felt like someone else's thinking. Not bad thinking. Just... not mine. The insights I usually have? Blunted. The analysis was correct but felt generic.
Week 2-3: Reality Check
I pulled back. Used AI only for the friction tasks. Research, email drafting, code scaffolding. Kept the thinking for myself.
This is crucial. AI is best at automating the work that doesn't require your unique perspective. Email? Generic. Data cleaning? Generic. Strategic analysis of why a pharma stock's valuation is contrarian? That needs you. Your experience. Your edge.
Week 4+: The Sustainable Version
I settled into a rhythm. Every morning, I batch my AI work. 20 minutes before my real work starts. I use AI to clear the desk: emails drafted, research done, datasets prepped. Then I close ChatGPT and Claude.
The actual thinking happens in focus blocks. No AI. Just me, the data, and thinking deeply.
I'd estimate the breakdown now: 70% of my work time is uninterrupted thinking. 20% is collaborative work (meetings, discussions, reviews). 10% is friction work (email, administrative).
Before? It was roughly 25% thinking, 40% friction, 35% collaborative. Those 45 percentage points? They're where the 10x lives.
| Work Type | Before AI | After AI | Weekly Hours Saved |
|---|---|---|---|
| Deep Thinking & Analysis | 10 hours | 28 hours | +18 |
| Friction Work (Email, Admin) | 16 hours | 4 hours | -12 |
| Meetings & Collaboration | 14 hours | 8 hours | -6 |
The Traps (Because There Are Several)
Trap 1: Using AI for Thinking, Not Friction
This is the biggest one. I see colleagues using Claude to write their entire analysis. Then spending time "reviewing" it. That's not productivity. That's procrastination with extra steps.
If you're spending more time editing AI output than it would've taken to write originally, you're doing it wrong. Use it only when it's faster AND you understand the output deeply enough to verify it.
Trap 2: The Perfectionism Tax
AI outputs are good enough for 80% of cases. But some part of you will want to iterate, tweak, refine. That tweaking eats the time you saved.
My rule: first draft from AI gets 5 minutes of editing max. If it's not right after that, I scrap it and do it myself. This prevents the endless loop of "just one more revision."
Trap 3: Treating It Like a Job Replacement
If you're worried about AI replacing you? Good. That means you understand the risk. But the risk isn't real if you're using AI as a tool to do better work, not to do less work.
The people AI will replace are the ones who use it to cut corners. The people who'll thrive are the ones who use it to go deeper.
My Perspective
Here's what surprised me: the productivity gain wasn't about working faster. It was about thinking better.
In my role at Morningstar, I've analyzed fund performance data for hundreds of portfolios. I used to spend so much mental energy on the *how* (Excel formulas, pivot tables, cleaning data) that the *why* got shallow. Why is this fund underperforming? I'd have surface-level answers.
Now, because the data prep is automated, I can ask better questions. I can run 5 different analyses instead of 1. I can challenge my assumptions. Last quarter, this led to a recommendation that saved the organization roughly ₹4 crores in a rebalancing decision. That wouldn't have happened if I was still spending 3 hours on data setup instead of 15 minutes.
What surprised me most? I'm less tired. I thought productivity meant doing more. Turns out, it means protecting your energy for things only you can do. AI is beautiful at that protection.
I got one thing wrong though: I thought I'd need to learn to code. I don't. Claude handles it. My mistake was assuming technical depth mattered more than understanding what questions to ask. It doesn't.
Final Thoughts
If you're in Kalyan, taking the 5:47 AM local to Mumbai, spending 8 hours in a cubicle, and coming home tired — this is for you.
You don't need to become an AI expert. You need to become ruthless about friction. Every task that doesn't require your unique thinking is a candidate for automation. Not because automation is trendy. But because your time is the most valuable resource you have, and you're trading it for something that doesn't compound.
Start small. Pick one tool. One friction point. Give it two weeks. If it works, expand. If it doesn't, move to the next tool.
The goal isn't to work 10x harder. It's to think 10x deeper. And for that, you need time. AI gives you time back.
You've got this. And if you're struggling with where to start, remember: the best tool is the one you'll actually use daily, not the one that sounds coolest at a tech meetup.
– 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 • 07 June 2026
.png)
.gif)
0 Comments