Finance & Data Analyst — zero to hero
From formula-googling and copy-paste reports to explaining any dataset in plain language — with verification instincts to match.
You're a hero when…
You interrogate spreadsheets in plain English, draft the narrative before the meeting, and catch AI's invented numbers before anyone else sees them.
13 steps · 📖 read a guide · 🛠️ try a tool · 💪 do a real mission (with a copyable prompt)
0 of 13 done
1 Foundations
- Step 1 📖 Read
What Is AI, Actually? →
Why a system that writes beautifully still can't reliably multiply — and what that means for trusting it with numbers.
- Step 2 📖 Read
Why AI Makes Things Up →
Non-negotiable for your role: AI states wrong numbers with perfect confidence. Build the verification reflex first.
- Step 3 📖 Read
Prompting Basics →
Analysis prompts live and die on context: what the data is, what decision it feeds, what "good" looks like.
- Step 4 🛠️ Try
Token Counter →
Know how much data you can actually paste before the model's memory overflows mid-analysis.
2 Daily reps
- Step 5 💪 Do
Interrogate the dataset
Your daily bread: paste data, ask what it means, verify what matters.
Show the mission prompt
Here's a data table: [paste]. First describe what the data IS (columns, period, units) so I can confirm you read it right. Then: 3 most important patterns, anything anomalous, and the one chart that would show the main story. For every number you cite, show the cells it came from.
- Step 6 💪 Do
Formula & query helper
Excel, Sheets, or SQL — describe the goal, get the syntax, understand it.
Show the mission prompt
In [Excel/Google Sheets/SQL], I have [describe tables/columns]. I need to [goal]. Give me the exact formula/query, a one-sentence explanation of how it works, and one edge case it might break on (blanks, duplicates, timezone) with the fix.
- Step 7 💪 Do
The forecast stress-test
Use AI as the skeptic, not the oracle.
Show the mission prompt
Here's my forecast and its assumptions: [paste]. Attack it: which assumption is most fragile, what would have to be true for the number to be 30% off, what seasonal or one-time effects might I be extrapolating? Then suggest the two sensitivity checks most worth running. Do NOT generate your own forecast.
- Step 8 💪 Do
Numbers → narrative
The report writes itself; you verify and own it.
Show the mission prompt
Turn this analysis into an executive summary: [paste key figures + findings]. Structure: the headline in one sentence, 3 supporting points each anchored to a number, one risk, one recommendation. No jargon, under 200 words. Mark any statement that's interpretation rather than fact, so I can gut-check it.
3 Power moves
- Step 9 🛠️ Try
System Prompt Architect →
An analyst assistant that knows your company's metrics, definitions, and reporting calendar — consistent analysis every time.
- Step 10 🛠️ Try
RAG Explorer →
See how "chat with your documents" works — the architecture behind querying policy docs and reports safely.
- Step 11 📖 Read
AI Privacy & Safety Basics →
Financials are the most sensitive thing you touch. Round numbers, strip names, know your company's approved tools.
4 Hero level
- Step 12 💪 Do
The monthly reporting pipeline
Capstone: your month-end, systematized.
Show the mission prompt
Help me template my monthly reporting. Interview me about my recurring reports, data sources, and audiences. For each report, build a reusable prompt with {{variables}} for the fresh numbers, plus a verification checklist of what I must manually confirm before sending. The goal: month-end in half the time with MORE accuracy, not less. - Step 13 📖 Read
Loop Engineering →
Where analysis is heading: agents that pull data, run checks, and draft reports — learn the loop before it learns your job.
🏆 Path complete!
You didn't just read about AI — you practiced it on your actual work. Keep the missions in your weekly routine, and consider a second path: the foundations carry over.