🚀 Claude AI Task-Based Guide: 8 Categories, 40+ Prompts, 12 Code Examples (Bangla Tutorial 2026) | FreeLearning365

 

Bangla AI Tutorial Series — Post 3 of 20

Claude AI দিয়েTask-Based কাজ

How to Use Claude Based on Task Categories

Same AI. Different results. The difference? Your task clarity.

Writing থেকে Coding, Data Analysis থেকে Business Communication — প্রতিটা কাজের জন্য exact prompts, real examples, এবং working code। এই একটা post পড়লে তুমি ৮০% কাজ Claude দিয়ে করতে পারবে।

8

Task Categories

40+

Real Prompts

12

Code Examples

20K+

Words of Content

🚀 পড়া শুরু করো📌 Real Scenarios দেখো

Scroll

🔥 Key Insight

Claude-এর সব power কি তুমি use করছো?

গবেষণায় দেখা গেছে বেশিরভাগ মানুষ Claude-এর মাত্র ২০% capability use করে। তারা শুধু simple questions করে, simple answers পায়। কিন্তু Claude আসলে পারে — code লিখতে, data analyze করতে, business decisions support করতে, creative content তৈরি করতে, complex research summarize করতে। এই post-এ তুমি শিখবে কোন কাজের জন্য কোন approach, কোন prompt structure, এবং কীভাবে প্রতিটা task category-তে maximum output বের করতে হয়।

Task Categories

৮টা Task Category — কোনটা তোমার?

Claude-এর capability বুঝতে হলে প্রথমে জানতে হবে — সে কোন ধরনের কাজে কতটা ভালো। নিচের ৮টা category-তে Claude exceptional।

✍️

Writing & Editing

Articles, emails, reports, social media, creative writing। Claude-এর সবচেয়ে strong suit।

💻

Coding & Dev

Code writing, debugging, review, documentation। HumanEval-এ 92% score।

📊

Data Analysis

Data interpretation, visualization planning, statistical insights, report writing।

💼

Business Comm.

Proposals, presentations, meeting summaries, strategy documents।

🔬

Research

Literature review, summarization, competitive analysis, fact synthesis।

🎨

Creative Content

Scripts, storytelling, brainstorming, campaign ideas, creative briefs।

📚

Learning & Teaching

Concept explanation, quiz generation, lesson plans, study guides।

⚙️

Automation Planning

Workflow design, process optimization, system thinking, SOPs।

✍️

Category 01

Writing & Editing

Claude-এর সবচেয়ে শক্তিশালী দিক হলো writing। Article, email, report, social media, creative writing — প্রতিটায় সে exceptional। কিন্তু prompt ঠিকমতো না দিলে output generic হয়।

Article WritingEmail DraftingReport WritingSocial MediaEditing & ProofreadingSEO Content

Writing-এ Claude কেন এত ভালো? কারণ সে context বোঝে — তুমি কাকে লিখছো, কেন লিখছো, কী effect চাও — এই সব একসাথে process করে।

Prompt Structure যা কাজ করে

⚠️ সবচেয়ে Common Mistake

বেশিরভাগ মানুষ বলে: "একটা article লিখো।" — এটা দিয়ে generic output আসে।
সঠিক approach: Role + Context + Audience + Format + Tone — সব একসাথে দাও।

✍️ Article Writing

Complete Article Package — ১ Prompt-এ সব

এই prompt structure দিলে Claude একটা complete article package দেবে — outline থেকে conclusion পর্যন্ত।

তুমি একজন experienced [NICHE] content writer। আমি [PLATFORM]-এর জন্য একটা article লিখতে চাই। Topic: [TOPIC] Target audience: [AUDIENCE DESCRIPTION — বয়স, location, interest] SEO keyword: [PRIMARY KEYWORD] Word count: [TARGET WORDS] Tone: [CONVERSATIONAL / FORMAL / EXPERT / FRIENDLY] Language: [BANGLA / ENGLISH / MIXED] আমাকে দাও: 1. H1 Title — 3 versions (emotional / SEO-optimized / curiosity-based) 2. Meta description (155 chars) — 2 versions 3. Detailed outline — H2 + 3-4 bullet points per section 4. Introduction — 150 words with strong hook 5. Body paragraphs — main sections fully written 6. Conclusion — 100 words with CTA 7. 10 relevant hashtags সব একসাথে, এক response-এ।

Claude-এর Response Example:

Title Options:
1. [Emotional] "আমি ৩ ঘণ্টার কাজ ২০ মিনিটে করি — এই AI Tool-টাই কারণ"
2. [SEO] "Claude AI দিয়ে Content Writing: Complete Beginner Guide 2025"
3. [Curiosity] "বাংলাদেশের Freelancer-রা যে AI Secret ব্যবহার করছে সেটা কী?"

Meta Description (v1): Claude AI দিয়ে content writing শিখুন। Real examples, step-by-step guide, বাংলায় সম্পূর্ণ tutorial। আজই শুরু করুন — বিনামূল্যে। [147 chars]

[Full article continues with all sections...]

📧 Email Writing

Real Scenario: Sumaya-র Client Email Problem

সুমাইয়া একটা digital agency-তে account manager। প্রতিদিন ৩০-৪০টা client email। প্রতিটায় different tone, different context।

❌ Wrong Prompt

"একটা client email লিখো"

Output: Generic, formal, lifeless

Context missing: কোন client? কী topic?

Tone missing: friendly? stern?

Goal missing: কী result চাও?

✅ Right Prompt

"আমি একজন agency account manager"

Client: Rahim Textile Ltd, 3 বছরের পুরনো client

Context: Project ২ সপ্তাহ late হচ্ছে

Goal: Apologize করো, trust রাখো, timeline দাও

Tone: Apologetic কিন্তু professional

আমি একটা digital agency-র account manager। আমার একটা client (Rahim Textile Ltd — ৩ বছরের পুরনো relationship) কে email করতে হবে। Situation: তাদের website redesign project ২ সপ্তাহ late হচ্ছে। Main reason: developer illness + scope creep। Goal: Apology করো, trust maintain করো, নতুন realistic timeline দাও। New deadline: ১৫ দিন পর। Offer: ৫% discount on final invoice as goodwill। Tone: Sincere apology, professional, relationship-preserving। Defensive না। Language: Formal English। Include: Subject line, full email body।

Subject: Update on Rahim Textile Website Project — Revised Timeline & Our Commitment

Dear Mr. Rahim,

I want to reach out personally to address the delay in your website redesign project, and to sincerely apologize for falling short of our committed deadline.

Due to an unexpected illness within our development team, combined with the scope enhancements we incorporated based on your valuable feedback last month, we are unable to deliver by the originally agreed date. I take full responsibility for not communicating this earlier, and I understand how this affects your planned launch timeline.

Here is our revised commitment:
• New delivery date: [Date — 15 days from today]
• Dedicated QA review: 3 additional testing rounds
• Daily progress updates until delivery
• Goodwill gesture: 5% reduction on your final invoice

Your trust in us over the past three years means everything to our team. We are committed to delivering a website that exceeds your expectations.

Could we schedule a brief call this week to walk you through the current progress?

Warm regards,
[Your Name]

🔧 Editing & Proofreading Prompts

Grammar + Style Check

এই text-টা review করো: [PASTE TEXT] দাও: 1. Grammar errors — সব 2. Awkward sentences 3. Tone consistency issues 4. Improved version — complete rewrite 5. কী change করলে, কেন

Bangla-English Mixed Polish

এই Bangla-English mixed text-টা professional করো: [PASTE TEXT] Rules: - Code-switching natural হবে - Bangla part grammatically correct - English part formal - Overall tone: [SPECIFY] Polished version দাও।

✅ Writing-এ Best Practices

• Role দাও: "তুমি একজন experienced journalist" — role দিলে Claude সেই perspective থেকে লেখে।
• Audience specify করো: "৪০+ বয়সী Bangladeshi business owner" — audience জানলে tone automatically adjust হয়।
• Example দাও: "এই style-এ লেখো: [EXAMPLE]" — একটা example ১০০০ word instruction-এর চেয়ে ভালো।
• Iterate করো: "এটা আরো formal করো" / "shorter করো" / "more emotional করো" — Claude iterate করতে পারে।

💻

Category 02

Coding & Development

Claude coding-এ exceptional। HumanEval benchmark-এ 92% — GPT-4o-এর সমান। Code লেখা, debugging, review, documentation — সব কিছুতে দক্ষ। সঠিক prompt দিলে production-ready code পাবে।

PythonJavaScriptSQLReactAPI DesignDebuggingCode Review

🐛 Bug Debugging

Real Scenario: Arif-এর Python Bug — ৩ ঘণ্টার সমস্যা ১৫ মিনিটে

আরিফ একজন junior developer। একটা FastAPI endpoint-এ authentication bug। ৩ ঘণ্টা ধরে Stack Overflow-এ খুঁজছে, কিছু পাচ্ছে না।

আমার FastAPI code-এ authentication bug আছে। JWT token verify করার সময় error হচ্ছে। Error message: jose.exceptions.JWTError: Signature verification failed Code: ```python from jose import JWTError, jwt from datetime import datetime, timedelta SECRET_KEY = "my-secret-key" ALGORITHM = "HS256" def create_token(data: dict): to_encode = data.copy() expire = datetime.utcnow() + timedelta(hours=24) to_encode.update({"exp": expire}) return jwt.encode(to_encode, SECRET_KEY, algorithm=ALGORITHM) def verify_token(token: str): try: payload = jwt.decode(token, SECRET_KEY, algorithms=ALGORITHM) return payload except JWTError: raise HTTPException(status_code=401, detail="Invalid token") ``` Framework: FastAPI | Library: python-jose কেন error হচ্ছে এবং fixed version দাও।

Python — Fixed Code by Claude

# Claude-এর Analysis: # Bug: algorithms parameter-এ list দেওয়া হয়নি — string দেওয়া হয়েছে # jwt.decode() এর algorithms param একটা list expect করে, string না। from jose import JWTError, jwt from datetime import datetime, timedelta from fastapi import HTTPException, status from typing import Optional import os # ✅ Fix 1: Secret key কখনো hardcode করো না SECRET_KEY = os.getenv("SECRET_KEY", "change-this-in-production") ALGORITHM = "HS256" ACCESS_TOKEN_EXPIRE_HOURS = 24 def create_access_token(data: dict, expires_delta: Optional[timedelta] = None): to_encode = data.copy() if expires_delta: expire = datetime.utcnow() + expires_delta else: expire = datetime.utcnow() + timedelta(hours=ACCESS_TOKEN_EXPIRE_HOURS) to_encode.update({"exp": expire, "iat": datetime.utcnow()}) encoded_jwt = jwt.encode(to_encode, SECRET_KEY, algorithm=ALGORITHM) return encoded_jwt def verify_token(token: str) -> dict: credentials_exception = HTTPException( status_code=status.HTTP_401_UNAUTHORIZED, detail="Could not validate credentials", headers={"WWW-Authenticate": "Bearer"}, ) try: # ✅ Fix 2: algorithms=[LIST] — not algorithms="STRING" payload = jwt.decode( token, SECRET_KEY, algorithms=[ALGORITHM] # ← এটাই Bug ছিল! ) # ✅ Fix 3: Token expiry check করো user_id: str = payload.get("sub") if user_id is None: raise credentials_exception return payload except JWTError as e: # ✅ Fix 4: Log the error (don't hide it) print(f"JWT verification failed: {e}") raise credentials_exception # Usage example: # token = create_access_token({"sub": "user_123", "role": "admin"}) # payload = verify_token(token) # → {"sub": "user_123", "role": "admin", "exp": ...}

Claude-এর Explanation (বাংলায়)

মূল bug: algorithms=ALGORITHM এখানে string দেওয়া হয়েছে, কিন্তু jwt.decode() function list expect করে। সঠিক হবে algorithms=[ALGORITHM]।

Additional fixes: Secret key hardcode না করা, token expiry check করা, এবং proper error logging।

🔍 Code Review

Production-Ready Code Review — React Component

তুমি একটা React component লিখেছো। Claude দিয়ে professional code review করাও — bugs, performance, accessibility সব একসাথে।

এই React component-টা professionally review করো। ```jsx function UserList() { const [users, setUsers] = useState([]); useEffect(() => { fetch('/api/users') .then(res => res.json()) .then(data => setUsers(data)); }, []); return (

{users.map(user => (

{user.name}

{user.email}

))}

); } ``` Review করো: 1. Bugs এবং potential issues 2. Performance problems 3. Error handling gaps 4. Accessibility issues 5. Best practices violations 6. Complete improved version

React JSX — Claude-এর Improved Version

import React, { useState, useEffect, useCallback } from 'react'; /** * Claude-এর Review Findings: * ❌ Bug 1: Error handling নেই — fetch fail হলে silent * ❌ Bug 2: Loading state নেই — UX খারাপ * ❌ Bug 3: Memory leak — component unmount হলে fetch cancel হয় না * ❌ Bug 4: Accessibility — semantic HTML নেই * ❌ Bug 5: Empty state নেই — no users হলে কী দেখাবে? */ function UserList() { const [users, setUsers] = useState([]); const [loading, setLoading] = useState(true); const [error, setError] = useState(null); const fetchUsers = useCallback(async (signal) => { try { setLoading(true); setError(null); const res = await fetch('/api/users', { signal }); if (!res.ok) { throw new Error(`HTTP error: ${res.status}`); } const data = await res.json(); setUsers(data); } catch (err) { if (err.name !== 'AbortError') { setError(err.message); } } finally { setLoading(false); } }, []); useEffect(() => { // ✅ AbortController — memory leak prevent করে const controller = new AbortController(); fetchUsers(controller.signal); return () => controller.abort(); }, [fetchUsers]); // ✅ Loading state if (loading) return ( <div role="status" aria-label="Loading users"> <p>Loading users...</p> </div> ); // ✅ Error state if (error) return ( <div role="alert"> <p>Error: {error}</p> <button onClick={() => fetchUsers()}>Retry</button> </div> ); // ✅ Empty state if (users.length === 0) return <p>No users found.</p>; return ( // ✅ Semantic HTML for accessibility <section aria-label="User List"> <h2>Users ({users.length})</h2> <ul> {users.map(user => ( <li key={user.id}> <strong>{user.name}</strong> <span>{user.email}</span> </li> ))} </ul> </section> ); }

🗃️ SQL Optimization

SQL Query Optimization — ১০ সেকেন্ড → ০.২ সেকেন্ড

Nadia একটা e-commerce company-র data analyst। একটা report query ১০ সেকেন্ড নিচ্ছে। Production database-এ এটা unacceptable।

এই SQL query টা optimize করো। ১০ সেকেন্ড নিচ্ছে — production-এ unacceptable। ```sql SELECT c.customer_name, c.email, COUNT(o.id) as total_orders, SUM(o.total_amount) as lifetime_value FROM customers c LEFT JOIN orders o ON c.id = o.customer_id WHERE o.created_at >= '2024-01-01' AND o.status = 'completed' GROUP BY c.id, c.customer_name, c.email ORDER BY lifetime_value DESC; ``` Table sizes: - customers: 500,000 rows - orders: 5,000,000 rows Database: PostgreSQL 15 Current execution time: ~10 seconds Target: < 0.5 seconds Optimized query + index recommendations দাও।

SQL — Claude-এর Optimized Version

-- Claude-এর Analysis: -- Problem 1: LEFT JOIN কিন্তু WHERE-এ orders filter → Implicit INNER JOIN -- Problem 2: created_at এ index নেই → full table scan -- Problem 3: status filter combined index নেই -- Problem 4: Subquery দিয়ে pre-filter করলে join আগে ছোট হবে -- STEP 1: Indexes তৈরি করো CREATE INDEX CONCURRENTLY idx_orders_date_status ON orders (created_at, status) WHERE status = 'completed'; -- Partial index! CREATE INDEX CONCURRENTLY idx_orders_customer_id ON orders (customer_id) INCLUDE (total_amount, created_at, status); -- STEP 2: Optimized Query WITH completed_orders AS ( -- Pre-filter orders BEFORE joining (dramatically reduces join size) SELECT customer_id, COUNT(*) AS total_orders, SUM(total_amount) AS lifetime_value FROM orders WHERE created_at >= '2024-01-01' AND status = 'completed' GROUP BY customer_id ) SELECT c.customer_name, c.email, co.total_orders, co.lifetime_value FROM completed_orders co -- INNER JOIN (not LEFT JOIN) — only customers who ordered INNER JOIN customers c ON c.id = co.customer_id ORDER BY co.lifetime_value DESC; -- Expected result after indexes: ~0.15-0.3 seconds -- Improvement: 30-60x faster -- STEP 3: Verify with EXPLAIN ANALYZE EXPLAIN (ANALYZE, BUFFERS, FORMAT TEXT) WITH completed_orders AS (...) SELECT ...;

✅ Coding Task-এ Best Practices

• Full context দাও: Language, framework, version — সব specify করো।
• Error পুরোটা paste করো: Half error message = half answer।
• Ask for explanation: "কোডটা explain করো" — শুধু code পেলে কিছু শিখবে না।
• Test cases চাও: "Unit tests দাও" — Claude production-ready tests লিখতে পারে।
• Iterate করো: "আরো efficient করো" / "Edge cases handle করো" — Claude improve করবে।

📊

Category 03

Data Analysis

Data থেকে insight বের করা, visualization plan করা, statistical summary লেখা, trend analysis — Claude এসব কাজে অসাধারণ। Actual computation না করলেও analysis framework দিতে পারে।

Sales AnalysisPython PandasExcel FormulaData VisualizationStatistical SummaryDashboard Design

📈 Sales Analysis

Real Scenario: 6 মাসের Sales Data Analysis — Karim-এর E-commerce Business

করিম ভাই-এর একটা online shop আছে। ৬ মাসের sales data আছে CSV-তে। Claude-কে analysis করতে দিলে — পুরো business insight বের হয়।

আমার e-commerce business-এর ৬ মাসের sales data আছে। Columns: Date, Product_Name, Category, Units_Sold, Unit_Price, Customer_Location, Customer_Age_Group Sample data: 2024-07-01, Kurti Set, Women's Fashion, 12, 1200, Dhaka, 25-35 2024-07-01, Punjabi, Men's Fashion, 8, 900, Chittagong, 35-45 2024-07-02, Sharee, Women's Fashion, 5, 3500, Dhaka, 45-55 ... আমাকে complete analysis করো: 1. Monthly revenue trend (কোন মাস সবচেয়ে ভালো?) 2. Top 5 products by revenue 3. Best performing location 4. Age group analysis 5. Category performance 6. Seasonal patterns 7. Business recommendations (specific, actionable) Python code দাও যেটা এই analysis automatically করবে।

Python — Complete Sales Analysis Script

import pandas as pd import numpy as np from datetime import datetime import warnings warnings.filterwarnings('ignore') class SalesAnalyzer: """Complete e-commerce sales analysis — Bangladeshi market.""" def __init__(self, filepath: str): self.df = self._load(filepath) def _load(self, filepath: str) -> pd.DataFrame: df = pd.read_csv(filepath) df['Date'] = pd.to_datetime(df['Date']) df['Revenue'] = df['Units_Sold'] * df['Unit_Price'] df['Month'] = df['Date'].dt.strftime('%B %Y') df['Month_Num'] = df['Date'].dt.month print(ff"✅ Loaded {len(df):,} records | Revenue: ৳{df['Revenue'].sum():,.0f}") return df def monthly_trend(self): """মাসিক revenue trend।""" m = self.df.groupby(['Month_Num','Month']).agg( Revenue=('Revenue','sum'), Orders=('Date','count'), Units=('Units_Sold','sum') ).reset_index().sort_values('Month_Num') m['Growth_%'] = m['Revenue'].pct_change() * 100 print("\n📅 MONTHLY REVENUE TREND") print("="*65) for _, r in m.iterrows(): bar = "█" * int(r['Revenue'] / m['Revenue'].max() * 30) growth = (ff"({r['Growth_%']:+.1f}%)" if not np.isnan(r['Growth_%']) else "(Base)") print(ff"{r['Month']:<15} ৳{r['Revenue']:>10,.0f} {bar} {growth}") return m def top_products(self, n=5): """Top performing products।""" p = self.df.groupby('Product_Name').agg( Revenue=('Revenue','sum'), Units=('Units_Sold','sum'), Avg_Price=('Unit_Price','mean') ).reset_index() p['Share_%'] = (p['Revenue'] / p['Revenue'].sum() * 100).round(1) top = p.nlargest(n, 'Revenue') print(ff"\n🏆 TOP {n} PRODUCTS") print("="*65) for i, (_, r) in enumerate(top.iterrows(), 1): print(ff"{i}. {r['Product_Name']:<20} ৳{r['Revenue']:>10,.0f} ({r['Share_%']}% share)") return top def location_analysis(self): """বিভাগ অনুযায়ী analysis।""" loc = self.df.groupby('Customer_Location').agg( Revenue=('Revenue','sum'), Orders=('Date','count'), Avg_Order=('Revenue','mean') ).reset_index().sort_values('Revenue', ascending=False) print("\n🗺️ LOCATION ANALYSIS") print("="*65) for _, r in loc.iterrows(): pct = r['Revenue'] / loc['Revenue'].sum() * 100 bar = "▓" * int(pct / 2) print(ff"{r['Customer_Location']:<15} ৳{r['Revenue']:>10,.0f} {pct:.1f}% {bar}") return loc def recommendations(self, monthly, products, locations): """Actionable business recommendations।""" best_month = monthly.loc[monthly['Revenue'].idxmax(), 'Month'] best_product = products.iloc[0]['Product_Name'] best_location = locations.iloc[0]['Customer_Location'] worst_month = monthly.loc[monthly['Revenue'].idxmin(), 'Month'] print("\n💡 BUSINESS RECOMMENDATIONS") print("="*65) print(ff"1. {best_product} সবচেয়ে বেশি বিক্রি — এর stock সবসময় sufficient রাখো।") print(ff"2. {best_location} সবচেয়ে বড় market — এখানে targeted ads চালাও।") print(ff"3. {best_month}-এ সবচেয়ে ভালো ছিল — পরের বছর ওই মাসে campaign prepare করো।") print(ff"4. {worst_month}-এ revenue কম — এই সময়ে promotional offer দাও।") print(f"5. Bottom 20% products নিয়ে review করো — discontinue বা promote।") # Run the analysis: analyzer = SalesAnalyzer('sales_data.csv') monthly = analyzer.monthly_trend() products = analyzer.top_products(5) locations = analyzer.location_analysis() analyzer.recommendations(monthly, products, locations)

📊 Excel Formulas

Excel Formula Generation — HR Dashboard

Technical background ছাড়াও Claude Excel formulas লিখে দেয়। Sumaya HR Manager Excel-এ employee analytics dashboard বানাচ্ছেন।

আমি একটা Excel HR dashboard বানাচ্ছি। এই formulas দরকার: Sheet structure: - Sheet1: "Employees" — columns: A=ID, B=Name, C=Department, D=Salary, E=Join_Date, F=Status - Sheet2: "Analytics" — এখানে dashboard Analytics sheet-এ এই formulas দরকার: 1. Active employee count (Status="Active") 2. Average salary per department 3. Employees joined in last 12 months 4. Highest paid employee name 5. Department with most employees 6. Total salary cost per month 7. Employees on probation (< 6 months tenure) সব formula দাও — Excel এবং Google Sheets উভয়ের জন্য।

Excel / Google Sheets Formulas:

1. Active Employee Count:
=COUNTIF(Employees!F:F,"Active")

2. Average Salary by Department (e.g., HR dept):
=AVERAGEIF(Employees!C:C,"HR",Employees!D:D)

3. Joined in Last 12 Months:
=COUNTIFS(Employees!E:E,">="&DATE(YEAR(TODAY()),MONTH(TODAY())-12,DAY(TODAY())),Employees!F:F,"Active")

4. Highest Paid Employee Name:
=INDEX(Employees!B:B,MATCH(MAX(Employees!D:D),Employees!D:D,0))

5. Department with Most Employees:
=INDEX(Employees!C:C,MATCH(MAX(COUNTIF(Employees!C2:C1000,Employees!C2:C1000)),COUNTIF(Employees!C2:C1000,Employees!C2:C1000),0)) [Ctrl+Shift+Enter for array]

6. Total Monthly Salary Cost:
=SUMIF(Employees!F:F,"Active",Employees!D:D)

7. Employees on Probation (<6 months):
=COUNTIFS(Employees!E:E,">="&(TODAY()-180),Employees!F:F,"Active")

💼

Category 04

Business Communication

Proposals, presentations, meeting summaries, strategy documents, stakeholder reports — business writing-এ Claude professional grade output দেয়। Context দিলে tone automatic হয়।

Business ProposalsMeeting MinutesPresentationsStrategy DocsStakeholder ReportsPress Releases

📋 Meeting Summary

Raw Notes → Professional Meeting Minutes — ৫ মিনিটে

Rahel একজন Project Manager। Meeting শেষে rough notes আছে। Claude-কে দিলে professional minutes পাবে।

আমি একজন Project Manager। এই meeting-এর raw notes থেকে professional minutes তৈরি করো। Meeting type: Q3 Product Roadmap Review Date: [Date] Attendees: CEO, CTO, Product Head, Dev Lead, Marketing Head Raw notes: - CEO: revenue কম হচ্ছে, faster delivery দরকার - CTO: team stretched, tech debt বড় problem - Product Head: user research দেখাচ্ছে mobile app দরকার - Dev Lead: mobile app 4 months লাগবে, current team দিয়ে না - Marketing: competitor X mobile app launch করছে next month - CEO: hire করবো নাকি outsource? CTO বললো outsource better এখন - Decision: mobile app outsource করবো, Q4-এ launch target - Action: CTO vendor shortlist করবে ২ সপ্তাহে, Product Head requirement document করবে - Next meeting: ২ সপ্তাহ পরে vendor review Format: Professional meeting minutes। Action items clearly highlighted।

MEETING MINUTES
Q3 Product Roadmap Review
Date: [Date] | Duration: [Time]
Attendees: [CEO Name], [CTO Name], [Product Head], [Dev Lead], [Marketing Head]
Prepared by: Rahel [PM] | Distribution: All attendees

EXECUTIVE SUMMARY
The meeting focused on addressing declining revenue through accelerated product delivery. The team aligned on outsourcing mobile app development for a Q4 launch, driven by competitive pressure and internal capacity constraints.

KEY DISCUSSION POINTS

1. Revenue & Delivery Urgency
CEO highlighted declining revenue metrics and stressed the critical need for faster product delivery to maintain market competitiveness.

2. Technical Capacity Assessment
CTO reported that the engineering team is currently at full capacity with significant technical debt. Internal mobile app development was assessed as unfeasible without additional resources (estimated 4-month timeline).

3. Competitive Intelligence
Marketing flagged that a key competitor (Company X) is launching a mobile application within the next month, creating urgent market pressure.

4. Build vs. Outsource Decision
After evaluating internal capacity, CTO recommended outsourcing as the optimal approach for speed and quality at the current stage.

✅ DECISIONS MADE
• Mobile app development will be outsourced to a third-party vendor
• Target launch: Q4 [Year]
• Internal team will focus on technical oversight and integration

📌 ACTION ITEMS
| Owner | Action | Deadline |
| CTO | Vendor shortlist (min. 3 vendors) with evaluation criteria | 2 weeks |
| Product Head | Complete mobile app requirement document | 2 weeks |
| All | Review vendor proposals and requirements doc | Next meeting |

NEXT MEETING: [Date — 2 weeks] | Focus: Vendor review & selection

📄 Business Proposal

Cold Outreach Proposal — Winning Format

Freelancer Nadia একটা potential client-এর কাছে proposal পাঠাবে। Claude দিয়ে winning proposal format বানাবে।

আমি একজন freelance web developer। একটা Bangladeshi fashion brand-এর কাছে proposal পাঠাবো। Client: Aarong-এর competitor, mid-size brand (Dhaka-based) Their pain point: পুরনো website, mobile-unfriendly, slow checkout My proposal: Full website redesign + e-commerce optimization Timeline: 6 weeks Budget: ৳2,00,000 Proposal লিখো: - Executive summary - Problem identification (client-specific) - Proposed solution - Timeline breakdown - Investment (price justification) - Why me (without sounding desperate) - Call to action Professional English, confident tone, 400-500 words।

WEBSITE TRANSFORMATION PROPOSAL
Prepared for: [Fashion Brand Name] | Prepared by: Nadia [Developer]

EXECUTIVE SUMMARY
This proposal outlines a comprehensive digital transformation for [Brand Name]'s online presence — addressing current friction points in user experience and checkout conversion to drive measurable revenue growth.

THE CHALLENGE
Your current website faces three critical issues impacting sales:
• 73% of your target customers browse on mobile — your site isn't optimized for them
• Average page load time of 6-8 seconds (industry standard: under 3 seconds)
• A checkout process requiring 7+ steps (leading to significant cart abandonment)
These aren't cosmetic issues — they directly reduce conversion rates by an estimated 40-60%.

PROPOSED SOLUTION
A complete platform redesign delivering: Mobile-first responsive design built for Bangladeshi shoppers | Optimized checkout (2-step process) with bKash/Nagad/COD integration | Page speed under 2 seconds | SEO-ready architecture for organic growth

TIMELINE — 6 WEEKS
Week 1-2: Discovery, wireframing, UI design approval
Week 3-4: Development & payment gateway integration
Week 5: Testing across devices and browsers
Week 6: Launch support, training, 30-day bug warranty

INVESTMENT: ৳2,00,000
This includes full design, development, payment integration, testing, and 30 days post-launch support. Comparable agency work in Dhaka: ৳3,50,000–5,00,000.

WHY THIS WORKS
I've delivered 3 Bangladeshi e-commerce platforms in the past 18 months, including [Reference]. Average client revenue increase: 35% in first quarter post-launch.

Ready to discuss? I'd welcome a 20-minute call this week.

✅ Business Communication Best Practices

• Audience specify করো: CEO-কে email আর junior staff-কে email — tone আলাদা।
• Desired outcome বলো: "এই email-এর goal হলো তাকে meeting-এ রাজি করানো।"
• Industry context দাও: "আমরা garments export company" — Claude সেই industry-র language use করে।
• Format specify করো: "Bullet points use করো" / "Paragraph format চাই" — structure control করো।

🔬

Category 05

Research & Summarization

Long document summarization, competitive analysis, literature review — Claude-এর 200K token context এখানে shine করে। একটা পুরো বই দাও, summary চাও।

Document SummarizationCompetitor AnalysisLiterature ReviewMarket ResearchFact Verification

Research Paper Summarizer

এই research paper-টা আমার জন্য summarize করো: [PASTE PAPER ABSTRACT + KEY SECTIONS] আমি: [YOUR PROFESSION/CONTEXT] দাও: 1. Main findings (3-5 bullets) 2. Methodology (1 paragraph) 3. Practical implications 4. Limitations 5. বাংলায় ১ paragraph summary (No jargon — plain language)

Competitive Analysis Framework

আমার business: [DESCRIBE] Competitors: [LIST 3-5] প্রতিটা competitor-এর জন্য: 1. Core product/service 2. Pricing strategy 3. Target audience 4. Key strengths 5. Weaknesses 6. USP (Unique Selling Point) তারপর: Market gap analysis করো — আমি কোথায় opportunity পাবো?

🔍 Market Research

Bangladesh E-Commerce Market Research — Complete Report

Bangladesh-এর fashion e-commerce market নিয়ে একটা comprehensive analysis করো। Context: আমি একটা online women's fashion brand start করতে চাই। Budget: ৳5 লাখ initial investment। Target: ১ বছরে profitable হতে চাই। Analyze করো: 1. Market size এবং growth rate (2024-25 data) 2. Key players এবং market share 3. Target consumer profile (demographics, behavior) 4. Most successful product categories 5. Pricing strategy analysis 6. Distribution channels (Daraz vs own website vs social) 7. Seasonal trends (Eid, Puja, winter) 8. Entry barriers এবং risks 9. My competitive positioning recommendation 10. First 90 days action plan Data-driven, specific, Bangladesh-context।

Bangladesh Women's Fashion E-Commerce: Market Analysis 2025

1. Market Size & Growth
Bangladesh's overall e-commerce market is estimated at $3.5-4.2B GMV (2024). Fashion/apparel represents approximately 28-32% of total e-commerce transactions — the largest single category. Annual growth rate: 22-28% YoY. Women's fashion specifically growing at ~35% — outpacing the broader market.

2. Competitive Landscape
• Daraz: Market leader (40%+ share) — your distribution channel, not competitor
• Shajgoj: Beauty + fashion, 15-20% women's share, strong community
• Aarong Online: Heritage brand, premium positioning, loyal customer base
• Facebook Commerce: Estimated 60%+ of small fashion brands sell primarily via FB
• Instagram Shops: Growing rapidly, especially 18-30 age group

3. Target Consumer Profile
Primary segment: Urban women 22-38, dual-income households, smartphone-native, price-conscious but quality-aware. Average fashion spend: ৳800-3,000 per purchase. Purchase frequency: 2-4 times/month. Discovery: 78% via social media, 15% via search, 7% direct.

4. Winning Product Categories
• Salwar kameez sets: Consistent bestseller year-round
• Kurti (casual + office): Price sweet spot ৳800-1,800
• Sharee (festive): Eid & Puja = 40% of annual sharee revenue
• Western casual: Growing 45% YoY in 22-30 age group

5. 90-Day Action Plan (৳5 Lakh Budget)
Month 1 (৳1.5L): Product sourcing, photography, brand identity, Instagram + Facebook setup
Month 2 (৳2L): Daraz shop launch + own website, Facebook/Instagram ads test
Month 3 (৳1.5L): Scale winning ad creatives, micro-influencer partnerships, first review collection
Break-even target: Month 4-5 with 35-40% gross margin maintained.

🎨

Category 06

Creative Content

YouTube scripts, podcast outlines, social media hooks, campaign ideas, storytelling — creative work-এ Claude একটা exceptional brainstorming partner। তোমার idea + Claude-এর structure = viral content।

YouTube ScriptsPodcast OutlinesSocial Media HooksCampaign IdeasStorytelling

🎬 YouTube Script

15-Minute YouTube Script — Hook থেকে CTA পর্যন্ত

Tasin-এর একটা tech YouTube channel আছে। ৩৫,০০০ subscribers। Claude দিয়ে complete script তৈরি করে — manual time ৩ ঘণ্টা থেকে ৪৫ মিনিটে নামিয়ে এনেছে।

আমার Bengali Tech YouTube channel-এ একটা video script লিখতে হবে। Channel: [CHANNEL NAME] | Subscribers: 35,000 Audience: বাংলাদেশি tech enthusiasts, 20-35 বছর Video topic: "Claude AI দিয়ে মাসে ৳50,000 আয় — সত্যি নাকি মিথ্যা?" Target length: 12-15 minutes Language: Bangla-English mixed (natural code-switching) Structure দরকার: 1. Hook (0:00-0:30) — first 30 seconds must be attention-grabbing 2. Pattern interrupt (2:00) — keep retention 3. 3 real case studies (with fake but realistic data) 4. Honest assessment (pros + cons + realistic timeline) 5. Step-by-step guide for beginners 6. End screen CTA Tone: বন্ধুর মতো — conversational, honest, no hype। [Timestamp] markers দিও।

[0:00-0:30] — HOOK
[On screen: "AI দিয়ে ৳50,000?!" text + suspicious emoji]

"Social media-তে প্রতিদিন দেখছো — 'AI দিয়ে লাখ টাকা আয়', 'ChatGPT দিয়ে passive income'। কিন্তু সত্যি কথা কী জানো? এই কথাগুলোর কিছুটা সত্যি, কিছুটা মিথ্যা, বেশিটা — misleading।
আজকে আমি তোমাকে সত্যি টা বলবো। Real cases দেখাবো। আর শেষে একটা practical roadmap দেবো।
[Point at camera] Stay করো শেষ পর্যন্ত।"

[0:30-2:00] — CONTEXT SETUP
"আগে বুঝি — AI দিয়ে income বলতে আসলে কী বোঝায়। মূলত ৩টা path আছে..."
[Graphics: 3 paths diagram]
Path 1: Existing skill + AI = faster delivery
Path 2: AI knowledge বেচা (courses, consulting)
Path 3: AI-powered services (automation, chatbots)

[2:00-2:15] — PATTERN INTERRUPT
[Sudden cut] "Wait — comment-এ বলো: তুমি কোন path-এ interested?"

[2:15-7:00] — CASE STUDIES
Case 1: Rafia (Content Writer)
"রাফিয়া Fiverr-এ content writing করে। ২০২৩ সালে income ছিল মাসে $150। Claude use শুরু করলো।
Month 1: $150 → $280 (output double হলো)
Month 6: $280 → $800 (better proposals + faster delivery)
Month 12: $800 → $1,800 (repeat clients + premium rates)
তার quote: 'Claude আমার assistant, boss না।'"

[Continues with Cases 2 & 3...]

[12:30-14:30] — STEP-BY-STEP FOR BEGINNERS
Step 1: claude.ai account খোলো (free)
Step 2: তোমার existing skill identify করো
Step 3: সেই skill-এ Claude কীভাবে help করতে পারে — test করো
Step 4: Fiverr/Upwork-এ ১টা gig তৈরি করো
Step 5: First month target: $50 — unrealistic goal রাখো না

[14:30-15:00] — CTA
"যদি আজকে ১টাই জিনিস মনে রাখো — AI তোমার skill replace করে না, multiply করে। Next video-তে দেখাবো কীভাবে Claude দিয়ে Fiverr gig optimize করবে। Subscribe button টা চেপে দাও — miss করবে না।"

Real World

Real Bangladeshi Scenarios

তত্ত্ব যথেষ্ট। এখন দেখো বাস্তবে কীভাবে বিভিন্ন profession-এর মানুষ Claude use করছে task category অনুযায়ী।

👩‍💼 HR Manager — Garments Company, Gazipur

Task CategoryWriting (Email)

Daily emails before Claude45 min writing time

Daily emails with Claude8 min writing time

Time saved daily37 minutes

Monthly time saved15+ hours

Net benefit৳5,000+ value/month

👨‍💻 Junior Developer — Startup, Dhaka

Task CategoryCoding (Debug)

Average bug fix time (before)3.5 hours

Average bug fix with Claude25 minutes

Speed improvement8x faster

Weekly hours saved12+ hours

Career impactSenior-level output ✅

📊 Business Analyst — NGO, Dhaka

Task CategoryResearch + Writing

Grant proposal (manual)5 days

Grant proposal with Claude1.5 days

Time saved per proposal3.5 days

Extra proposals per month2 additional

Potential grant impact$20,000+ per year 🔥

🎬 YouTuber — Tech Channel, Chittagong

Task CategoryCreative (Scripts)

Script time before Claude3 hours/video

Script time with Claude45 minutes/video

Extra videos per week+2 videos

Extra monthly views+80,000 views

Extra AdSense income$240/month 🚀

The Framework

Universal Prompt Framework

যেকোনো task category-তে এই framework use করলে output ৩x better হবে। মনে রাখো: RCTFE.

R

Role — Claude-কে একটা role দাও

"তুমি একজন experienced Bengali content strategist" বা "তুমি একজন senior Python developer with 10 years experience।" Role দিলে Claude সেই perspective থেকে respond করে। Generic AI response আর আসে না।

C

Context — Situation explain করো

"আমি একটা Dhaka-based garments company-র HR Manager। আমাদের ৫০০ employees।" Context ছাড়া Claude general answer দেবে। Context দিলে specific, applicable answer পাবে।

T

Task — Exactly কী চাও বলো

"একটা professional warning letter লিখো" — ভালো। "একটা professional warning letter লিখো যেটা: (1) empathetic হবে, (2) company policy mention করবে, (3) improvement path দেখাবে, (4) formal কিন্তু harsh না" — অনেক ভালো।

F

Format — Output format specify করো

"Bullet points use করো", "৩০০ words max", "Table format-এ দাও", "JSON format-এ output চাই", "বাংলায় লেখো" — format বলে দিলে Claude সেটা follow করে। Edit করার সময় কমে।

E

Example — একটা example দাও (optional but powerful)

"এই style-এ লেখো: [EXAMPLE]" — একটা example ১০০০ words of instruction-এর চেয়ে ভালো। Claude pattern বুঝে সেই style-এ লেখে। Brand voice maintain করতে example অপরিহার্য।

💡 RCTFE Example — একসাথে দেখো

[R] তুমি একজন experienced HR professional।
[C] আমি একটা 200-person garments factory-র HR Manager। একজন employee তৃতীয়বার late এসেছে।
[T] একটা formal warning letter লিখো — empathetic কিন্তু serious।
[F] Subject line + formal letter format, ১৫০-২০০ words।
[E] এই tone-এ: [PASTE EXAMPLE LETTER]

Honest Assessment

Claude-এর Pros & Cons — কোন Task-এ কতটা ভালো?

Task Category	Claude-এর Strength	Limitation	Verdict
✍️ Writing	Context-aware, nuanced, multiple styles	Needs your voice/personality added	⭐⭐⭐⭐⭐ Excellent
💻 Coding	Multi-language, debug, review, docs	Can't run code — verify output	⭐⭐⭐⭐⭐ Excellent
📊 Data Analysis	Interpretation, visualization planning	No live data — works with what you give	⭐⭐⭐⭐ Very Good
💼 Business Comm.	Professional tone, structure, clarity	Needs industry context to be specific	⭐⭐⭐⭐⭐ Excellent
🔬 Research	Synthesis, summary, analysis of provided text	Knowledge cutoff — no real-time data	⭐⭐⭐⭐ Good (verify facts)
🎨 Creative	Brainstorming, script structure, variations	Needs your unique voice & personality	⭐⭐⭐⭐ Very Good
📚 Teaching	Explains complex concepts simply, patient	May oversimplify advanced topics	⭐⭐⭐⭐⭐ Excellent
⚙️ Automation	Workflow design, process optimization, SOPs	Can't execute — only plans	⭐⭐⭐ Good (planning only)

✅ Claude কখন সেরা?

Clear, specific prompt দিলে — RCTFE framework use করলে।

Context-rich situation — background information থাকলে।

Iterative work — back-and-forth conversation-এ।

Long-form content — বড় documents, complex analysis।

Multilingual work — বাংলা-ইংরেজি mixed content।

⚠️ Claude-এর সীমাবদ্ধতা মনে রাখো

Real-time internet access নেই free plan-এ — fresh news পাবে না।

Hallucination হতে পারে — important facts verify করো।

Code run করতে পারে না — output test করে নাও।

Sensitive/confidential data দিও না।

Clinical/legal/financial advice-এ human expert-এর পরামর্শ নাও।

Alternatives

Claude vs Others — Task-Based Comparison

Task	Claude	ChatGPT	Gemini	Best Choice
Long Writing	⭐⭐⭐⭐⭐	⭐⭐⭐⭐	⭐⭐⭐	Claude
Coding (Python/JS)	⭐⭐⭐⭐⭐	⭐⭐⭐⭐⭐	⭐⭐⭐⭐	Tie
Image Generation	❌	⭐⭐⭐⭐⭐	⭐⭐⭐	ChatGPT Plus
Real-time Web Search	Paid only	Paid only	Free ✅	Gemini/Perplexity
Bangla Quality	⭐⭐⭐⭐⭐	⭐⭐⭐⭐	⭐⭐⭐	Claude
Document Analysis	⭐⭐⭐⭐⭐ (200K)	⭐⭐⭐⭐ (128K)	⭐⭐⭐⭐⭐ (1M)	Claude/Gemini
Business Writing	⭐⭐⭐⭐⭐	⭐⭐⭐⭐	⭐⭐⭐	Claude

Resources

References & Learn More

📖

Anthropic Docs

docs.anthropic.com — Official Claude documentation, API guides, best practices।

🔬

Anthropic Research

anthropic.com/research — Constitutional AI, Claude model papers, safety research।

📰

The Verge AI

theverge.com/ai — Industry news, Claude updates, AI tool comparisons।

🏆

LMSYS Arena

lmsys.org — Independent AI benchmark — Claude vs GPT vs Gemini real comparisons।

🇧🇩

BASIS Bangladesh

basis.org.bd — Bangladesh IT industry data, AI adoption statistics।

🎓

FreeLearning365.com

আমাদের website-এ পুরো 20-post series, detailed tutorials, prompt library।

#ClaudeAI#AIBangla#কৃত্রিম_বুদ্ধিমত্তা#AIForWork#BangladeshTech#ProductivityHack#PromptEngineering#AIWorkflow#TaskAutomation#WorkSmart#AIUseCases#FreeLearning365#AIটিউটোরিয়াল#DigitalBangladesh#ContentCreatorBD#DeveloperBD#FreelanceBangladesh#LearnAI#SmartWork#AISkills

🔥 Next Post

Post 4 আসছে: Claude AI for Professionals

Developer, HR, Analyst, Manager — প্রতিটা profession-এর জন্য exact workflows, prompts, এবং real industry examples। Miss করো না।

🌐 FreeLearning365.com↑ Top-এ যাও

FreeLearning365.com

Bangla AI Tutorial Series — Post 3 of 20 | Claude AI Task Category Guide

AnthropicClaude AIAPI DocsFreeLearning365

© 2025 FreeLearning365.com · Bangla AI Tutorial Series · All rights reserved · Powered by Claude AI