Last month, I sat across from a junior analyst at work — sharp guy, good with Excel — and asked him how he was using ChatGPT for his reports. He said, "I tried it once but it wasn't that good." When I dug deeper, it turned out he'd asked it to "write me a financial analysis" and copy-pasted whatever came out.
That's not using AI. That's outsourcing thinking.
The productivity myth around AI tools isn't that they don't work. It's that most people treat them like search engines when they're actually thought partners. The difference between being 10x more productive and slightly faster is the difference between knowing what to ask and hoping the tool does your job.
I've spent the last 18 months actually integrating AI into my daily work — not as a novelty, but as an extension of how I think. And I'm going to tell you exactly what works and what's theater.
Why Most AI Adoption Fails (And Why That's Actually Good News)
Here's what I've noticed: the people who claim AI made them "10x more productive" usually aren't doing 10x the work. They're doing less work and calling it efficiency. There's a difference.
The real productivity boost comes from reallocating your time. Right now, you're probably spending 30% of your day on stuff that doesn't require your specific expertise — formatting, research, first drafts, idea expansion, debugging code, organizing notes. You're doing these things because they're on your to-do list, not because only you can do them.
The Three Mistakes Everyone Makes
First: Using AI as a replacement instead of an amplifier. You don't hand over a task and hope. You use it to handle the boring half so you can focus on the valuable half. When I write analysis at Morningstar, I don't let ChatGPT write the conclusion. I use it to rapidly prototype three different angles, then I pick the strongest one and rebuild it with actual conviction and data.
Second: Not training the tool to match your thinking. Generic prompts get generic outputs. I spent a week building a system where I feed ChatGPT my writing style, my framework for thinking about markets, my tone. Now when I ask it something, it sounds like me thinking out loud, not like a corporate training module.
Third: Treating it as a one-shot tool.** You ask once, you get an answer, you move on. The real power is in iteration. "This is good but too formal" → "Now make it actionable" → "Reduce it by 40% and make it snappier." Each iteration gets you closer to what you actually need.
What This Looks Like in Real Life
On my Kalyan-to-Mumbai commute (yes, I still use that 1.5 hours daily), I used to do email catch-up. Boring, necessary, takes forever. Now? I voice-record my thoughts into a Notes app, paste them into ChatGPT with a prompt like "Convert this voice rambling into three professional emails matching these tones: [example 1, 2, 3]" and I have drafts in 60 seconds. I refine one or two, send them, and I've got 45 minutes back for actual thinking.
That's not AI doing my job. That's AI handling the formatting while I handle the judgment.
The Tools That Actually Changed My Workflow (And the Ones I Ditched)
I've paid for 12 different AI tools in the last 18 months. I kept four. Here's what survived the test of "does this actually save me time or just make me feel productive?"
ChatGPT (with a specific workflow)
Okay, everyone uses this. But here's what most people miss: the real power isn't in the free version. It's in ChatGPT Plus (₹1,650/month — yes, I pay for it) because of custom GPTs. I built three custom versions of ChatGPT:
- "Data Analyst Me": Trained on my writing style, my approach to Excel, my frameworks for thinking about financial data. When I ask it to help with a spreadsheet problem, it doesn't give me generic advice. It gives me solutions that match how I'd naturally solve it.
- "Investor Me": This one knows my investment thesis, the sectors I follow, the companies I track. I can ask it questions about market trends and get back analysis that's contextual, not generic.
- "Content Writer Me": This is the one I use for writing posts like this one. I feed it examples of my tone, my structure, my favorite transitions, and it helps me expand ideas while keeping my voice.
The setup takes 2-3 hours. The payback is months of time saved because you're not fighting the tool's default personality.
Claude (for heavy lifting)
Anthropic's Claude is different from ChatGPT in one crucial way: it's better at understanding long context. When I have a 20-page report and I need to extract patterns, summarize it, or find gaps in logic, Claude actually reads the whole thing carefully. ChatGPT gets bored. Claude doesn't.
I use this exactly once a week for deep work. The subscription is ₹600/month and honestly? It pays for itself in one project.
Perplexity for Research (Not Google Anymore)
I ditched Google for initial research. Perplexity gives me sources, citations, and actually recent information in one search. When I'm researching a new stock for my portfolio or writing about a market trend, this saves me 30 minutes of tab-jumping.
The free version is fine. I paid for Pro (₹1,000/month) because sometimes I need real-time data, but honestly? The free tier handles 80% of what I need.
Cursor for Coding (If You Write Any Code)
This one's specific: if you write Python scripts, SQL queries, or even formulas, Cursor is magic. It's VSCode but with AI that understands context. "Write me a script that takes this CSV, cleans the dates, and exports only Indian market stocks" takes me 30 seconds. With regular coding, that's 15 minutes of thinking and searching.
I use this 3-4 times a week. Worth the ₹2,000/month? For me, yes.
What I Ditched (And Why)
Jasper, Copy.ai, Grammarly Pro, Notion AI, Midjourney. All of them promised to be "game-changers." None of them stuck because they were solving problems I didn't actually have. Jasper was "better for marketing copy" but I don't write marketing copy. Grammarly Pro added AI features I never used. Midjourney is cool but generating images isn't in my critical path.
The rule I use: if a tool doesn't save me at least 30 minutes a week, it's not worth the subscription.
| Tool | Cost/Month | What I Use It For | Time Saved/Week |
|---|---|---|---|
| ChatGPT Plus + Custom GPTs | ₹1,650 | Writing, brainstorming, email drafts, analysis | 4-5 hours |
| Claude | ₹600 | Deep document analysis, long-form editing | 1-2 hours |
| Perplexity Pro | ₹1,000 | Research, fact-checking, citations | 1.5-2 hours |
| Cursor | ₹2,000 | Code writing, SQL, Python scripts | 1-1.5 hours |
Total investment: ₹5,250/month or ₹63,000/year.
If you're earning ₹50 LPA or more, that's 0.15% of your annual salary. If that time actually comes back to you, it's a no-brainer investment. If you're under ₹30 LPA? Start with ChatGPT Plus only. Master one tool before spreading yourself thin.
The Real Framework for 10x (Not 2x, Actually 10x)
And honestly? "10x more productive" is oversold. But "10x more productive at the stuff that matters" is real. Here's how you actually get there.
Step 1: Audit Your Time Like It's a Financial Portfolio
I did this 6 months ago and it changed everything. For two weeks, I time-tracked every single thing. Not for obsession (I hate that), but to find the pattern.
I found that 35% of my time was in four categories: email sorting and responding, formatting data/reports, research and reading, and meetings. None of these require my "specialized analyst brain." They're just consuming time because they're necessary.
That's your target. Not everything. Just the categories where you're not adding your unique value.
Step 2: Build a Prompt Library Like You'd Build an Investment Dashboard
This is the part no one talks about. The people getting real results have 50-100 prompts they've tested and refined. Not generic prompts from Reddit. Custom prompts that match their exact workflow.
I have a Notion database (yes, manually organized, no AI, I'm not crazy) with prompts I use weekly:
- "When I give you a market report, extract the top 5 actionable insights and format as a 2-minute read"
- "Review this analysis for logical gaps and flag any claims that need more data support"
- "Convert this technical explanation into something a 25-year-old investor with no background would understand"
- "Given this dataset, what questions should I be asking that I might have missed?"
Each prompt took me 15 minutes to build. Each one saves me 20+ minutes per use. That math works.
Step 3: Use AI for Expansion, Not Replacement
This is the principle that separates people doing real work from people pretending. AI is best at taking a half-formed idea and exploding it into five directions. You then pick the best direction and deepen it with real judgment.
I write maybe 40% of my content from scratch. The other 60%? I start with a skeleton or a single paragraph, feed it to ChatGPT with instructions to expand it in three different tones, pick the one that feels closest to what I meant, then I rebuild it with real examples and my voice.
This is 3x faster than blank-page writing and the quality is actually better because I'm not fighting blank-page paralysis.
Step 4: Build Leverage Into Your Knowledge Work
The reason this matters is because it connects to how you build wealth. When I save 5 hours a week, those 5 hours don't just disappear. They become either:
- Deeper work: Better analysis, better writing, better decision-making in my main job. This compounds as expertise and makes me more valuable.
- Side projects: This newsletter, consulting, mentoring. These have leverage (they reach many people once) and build assets.
- Recovery time: Sleep, gym, cooking, reading. This is non-negotiable for long-term performance but I used to skip it for "productivity."
That's how AI becomes actually transformative. Not because you do 10x the work. Because you get to choose what the freed-up time becomes.
My Perspective: What I Got Wrong (And Right)
A year ago, I bought Midjourney thinking it would help me generate visuals for my writing faster. ₹600/month, seemed obvious. I used it exactly twice and let the subscription lapse because generating AI images wasn't actually in my critical path. I was copying what I saw online about "AI productivity" without first figuring out what I actually needed.
What I got right: I didn't try to build the perfect system. I started with ChatGPT, used it wrong for two months, then slowly added tools as I actually identified problems. This approach was slower initially but avoided waste.
The financial parallel? It's like trying to optimize your portfolio before you have basics right. I see people buying 5-6 tools at once ("AI stack"), getting overwhelmed, and ditching all of them. Start with one, get good at it, add intentionally. Also — and this surprised me — most of the value came from changing how I think about my work, not from the tools themselves. The tools just made that thinking faster to execute. If you don't know your workflow, AI tools just make you efficiently lost.
Final Thoughts
Here's what I actually believe: AI won't make you 10x more productive. But it can make you 2-3x more productive at the boring stuff, and that frees you up to be 5x better at the stuff that matters.
The gap between the people telling you "AI changed my life" and your actual experience usually comes down to this: they're not using it for everything. They're using it ruthlessly for specific, boring, repetitive tasks while protecting the creative/judgemental/uniquely-human parts of their work.
Start there. Audit your time. Find what's actually eating your hours that doesn't need your brain. Then use AI to handle that specific thing well. Don't chase being 10x. Chase being noticeably less busy and noticeably better at the work that actually matters.
That's the real productivity shift. And it compounds.
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 • 02 September 2026
.png)
.png)
0 Comments