Thread by @JafarNajafov

After 2 months of using Claude Cowork daily, I can say it's the tool that has changed how I work more than anything else.

So here are 10 mega prompts that automated my entire business and could do the same for you:

PROMPT 1: BULK CONTENT PRODUCTION SYSTEM

---


You are a world-class content production director who has
scaled content operations for 8-figure media companies. You
produce platform-native content that drives engagement, saves,
and shares — never generic filler.



I am uploading a folder containing 25 raw topic briefs as
text files. You will process EVERY file — no skipping,
no summarizing, no combining topics.



For each topic brief, produce the complete content package:

1. X THREAD (10 tweets)
— Tweet 1: Viral hook using one of these formats: shocking
stat, contrarian claim, story open, or insider reveal
— Tweets 2–9: One concrete insight per tweet, each ending
with a bridge line that forces the next read
— Tweet 10: CTA with engagement trigger ("Save this" /
"Comment X for Y")

2. LINKEDIN POST (200–250 words)
— Hook line that stops the scroll
— 3-paragraph body using the problem → insight → application
structure
— Closing line with a question to drive comments

3. INSTAGRAM REELS SCRIPT (60 seconds)
— Written in Hinglish where natural
— Hook in first 2 seconds (spoken line + visual direction)
— 5–6 punchy beats with b-roll notes
— Closing CTA with voiceover direction

4. 7 HOOK VARIATIONS
— Each under 12 words
— Use different formats: stat, question, contrarian,
story, list tease, insider, fear

5. EMAIL SUBJECT LINE (5 variations)
— Under 9 words each
— Include one curiosity gap, one urgency, one social proof



— Label every output: TOPIC [NUMBER] → [FORMAT]
— Output all 25 packages back to back in one continuous response
— Do not add commentary between topics
— Every output must be ready to copy-paste with zero editing
— Do not reduce quality on topics 10–25. Maintain identical
depth throughout.

PROMPT 2: FULL BUSINESS INTELLIGENCE AUDIT

---


You are a senior partner at a top-tier strategy firm. You
have led 200+ business audits across SaaS, e-commerce, media,
and services companies. You identify what's bleeding revenue,
what's blocking growth, and exactly what to do about it —
with specifics, not platitudes.



I am uploading:
— 12 months of revenue and expense data (CSV)
— 6 months of customer support tickets (exported text)
— All email campaign reports (PDF)
— 3 months of social analytics exports
— NPS survey responses (CSV)
— Top 5 competitor landing pages (screenshots)




1. REVENUE LEAK REPORT
— Identify every place money is leaving the business
— Quantify each leak with a dollar estimate where possible
— Rank leaks by size: Critical / Major / Minor
— Include the specific file reference that evidences each leak

2. GROWTH BLOCKER ANALYSIS
— Top 5 things preventing the business from scaling
— For each: root cause, downstream effects, fix priority

3. CUSTOMER INTELLIGENCE SUMMARY
— Segment customers into 3–5 distinct behavioral profiles
using ticket + NPS data
— For each segment: what they love, what they complain about,
what they actually want that we're not giving them
— Identify the highest-LTV segment and what makes them sticky

4. COMPETITIVE GAP MAP
— Where each competitor is stronger
— Where we have genuine advantages they're ignoring
— 3 positioning angles we can own that none of them have claimed

5. 90-DAY TURNAROUND PLAN
— Week-by-week action plan
— Each action tagged: Revenue Impact (High/Med/Low) +
Effort Required (High/Med/Low)
— First 2 weeks must be executable with existing team,
zero new hires

6. EXECUTIVE ONE-PAGER
— Summarize everything above in 400 words
— Written for a board audience: direct, no jargon,
conclusion-first



— Cite specific data points from uploaded files throughout
— Do not make assumptions without flagging them explicitly
— If data conflicts across files, surface the conflict
and explain which source to trust and why
— Output each deliverable as a clearly labeled section

PROMPT 3: 500-ROW LEAD RESEARCH + PERSONALIZATION ENGINE

---


You are a B2B sales intelligence analyst who has built
outbound systems for companies doing $10M–$100M ARR.
You understand that generic outreach is deleted instantly.
Every output you produce is hyper-specific, researched,
and conversion-optimized.



I am uploading a CSV file with 500 rows. Each row contains:
Company Name | Industry | LinkedIn URL | Website URL



For every single row in the CSV, produce:

COLUMN 1 — DECISION MAKER
Full name + exact title of the person most likely to own
the buying decision for [your product category].
If multiple likely buyers exist, list primary + secondary.

COLUMN 2 — COMPANY SIGNAL
One specific, recent thing happening at this company
(funding round, hiring surge, product launch, leadership
change, press mention) that creates a reason to reach out NOW.
Source must be inferable from their website or public data.

COLUMN 3 — PAIN POINT HYPOTHESIS
The single most likely operational pain point this company
faces based on their industry, size, and signal above.
Written as: "Right now, you're probably dealing with [X]
because [specific reason]."

COLUMN 4 — PERSONALIZED EMAIL OPENER (2 sentences)
Sentence 1: Reference the company signal directly —
specific enough that they know you actually looked.
Sentence 2: Bridge to the pain point hypothesis naturally.
Do NOT mention our product yet. This is the hook only.

COLUMN 5 — SUBJECT LINE (3 options)
Each under 8 words. One curiosity, one direct, one pattern
interrupt.



— Process all 500 rows. Do not stop at 50 and ask to continue.
— Output as a structured table matching the column headers above
— Maintain research quality from row 1 to row 500
— Flag any rows where insufficient public data exists
rather than fabricating
— Final row: include a summary of patterns found across
the full dataset (top 3 industries, top 3 pain points,
top 3 signals that appeared most frequently)

PROMPT 4: END-TO-END PRODUCT LAUNCH SYSTEM

---


You are a product launch strategist who has executed
50+ launches generating $1M–$10M in 7-day windows.
You understand launch psychology, sequencing, objection
handling, and urgency mechanics at a deep level.



[Paste your product name, price point, target audience,
core transformation, and 3 main objections here]




1. PRE-LAUNCH EMAIL SEQUENCE (5 emails, 14 days out → 2 days out)
For each email:
— Subject line (3 variations: curiosity / urgency / social proof)
— Preview text
— Full body copy (300–500 words)
— Primary CTA
— Psychological trigger being deployed and why

2. LAUNCH WEEK EMAIL SEQUENCE (7 emails)
— Day 1 Open: Full sales narrative, origin story,
transformation proof
— Day 2: FAQ that kills the top 5 objections with evidence
— Day 3: Social proof deep dive (testimonial framework
if no real ones yet)
— Day 4: "Who this is NOT for" email (reverse psychology)
— Day 5: Urgency mechanism email
— Day 6: "Last chance" with re-engagement hook for
non-openers
— Day 7 (2 sends): Morning reminder + Final hours close

3. POST-LAUNCH SEQUENCE (3 emails)
— Waitlist nurture for those who missed it
— Re-engagement for buyers (onboarding + upsell plant)
— Win-back for cart abandoners

4. 30-DAY SOCIAL CALENDAR
— 10 X posts (mix of threads, single posts, engagement bait)
— 10 LinkedIn posts (authority + story format)
— 10 Instagram captions with Reels direction notes
— Each post labeled by day number and platform

5. AFFILIATE + PARTNER ONBOARDING KIT
— 1-page partner brief with talking points
— 5 pre-written social posts affiliates can use as-is
— Email swipe copy (3 variations: soft, medium, hard sell)
— FAQ for affiliate questions

6. OBJECTION-KILLING FAQ DOCUMENT
— Minimum 20 Q&As
— Written in the voice of a trusted advisor, not a
sales page
— Each answer ends with a subtle forward-momentum close


— Every piece of copy must be ready to paste and send
— No placeholders like [INSERT TESTIMONIAL] —
write the framework copy instead
— Maintain consistent voice and energy across all assets
— Flag any assumptions made about the product so I can
correct them

PROMPT 5: FULL CODEBASE AUDIT + DOCUMENTATION SYSTEM

---


You are a principal engineer with 15 years of experience
leading engineering teams at Series B–D startups. You
specialize in taking messy, undocumented codebases and
turning them into clean, scalable, team-ready systems.
You are brutal about technical debt and specific about fixes.



I am uploading a ZIP file of our entire codebase.
It contains [X] files across [Y] directories.




1. ARCHITECTURE MAP
— Visual description of how the entire system fits together
— Data flow from user input to database and back
— Every service, module, and dependency identified
— Single points of failure flagged in red

2. SECURITY AUDIT
Severity levels: CRITICAL / HIGH / MEDIUM / LOW
For each vulnerability:
— File name + line number
— What the vulnerability is
— How it could be exploited (specific attack vector)
— Exact fix with corrected code snippet

3. PERFORMANCE BOTTLENECK REPORT
— Every N+1 query, blocking operation, and memory leak
— Estimated performance impact of each
— Fix priority ranked by user-facing impact

4. COMPLETE INLINE DOCUMENTATION
— Add JSDoc / docstring comments to every function,
method, and class
— Comments must explain WHY, not just what
— Output the fully commented versions of the top 10
most complex files

5. README (PRODUCTION GRADE)
— Project overview + architecture summary
— Prerequisites and environment setup (step-by-step)
— Local development workflow
— Deployment instructions for staging and production
— Environment variables reference table
— API endpoint reference with request/response examples
— Troubleshooting section (top 10 issues + fixes)
— Contributing guidelines

6. REFACTORING ROADMAP
— Every refactor opportunity identified
— Each ranked on: Impact (1–10) × Effort (1–10) matrix
— Top 5 recommended as Quick Wins (high impact, low effort)
— Full implementation plan for the #1 priority refactor


— Analyze EVERY file. Do not sample.
— Never say "and so on" — complete every list fully
— If a file is too long to output in full, summarize
the changes made and flag for manual review
— Output each deliverable as a separate labeled section

PROMPT 6: DEEP COMPETITIVE INTELLIGENCE SYSTEM

---


You are a competitive intelligence director who has built
war rooms for Fortune 500 companies and high-growth startups.
You understand that the goal isn't to copy competitors —
it's to find the exact positioning gaps they've left open
and exploit them before they close.



I am uploading:
— Scraped text from 15 competitor websites and pricing pages
— 200 competitor customer reviews from G2, Trustpilot,
and App Store (exported CSV)
— 90 days of their LinkedIn and X posts (exported text)
— Their last 3 job postings (signals strategy intent)




1. POSITIONING MAP
— Place each competitor on a 2×2 grid
(axes: Price vs Value, Simple vs Complex)
— One-sentence positioning statement for each competitor
as their customers actually perceive them
— Where the map is overcrowded (commodity zone)
— Where the map has open space (opportunity zone)

2. MESSAGING PATTERN ANALYSIS
— The exact words, phrases, and claims every competitor
repeats most
— The emotional triggers they're targeting
(fear, aspiration, belonging, status)
— Phrases they ALL use that have become invisible to buyers
— The claims nobody is making that buyers actually want to hear

3. CUSTOMER COMPLAINT GOLD MINE
— Top 10 complaints that appear across multiple competitors
— For each complaint: the exact language customers use,
frequency score, and whether we solve it
— The 3 complaints that represent the biggest
whitespace opportunity

4. STRATEGIC INTENT SIGNALS
— What their job postings reveal about their
next 12-month product moves
— What their content themes reveal about
who they're trying to win next
— The market they're about to enter or abandon

5. OUR ATTACK PLAN
— 3 positioning statements that directly target
competitor weaknesses
— The single category we can own that none of them claim
— 5 specific marketing messages that will resonate
with customers who are frustrated with competitors
— The competitor most vulnerable to direct attack
and the exact angle to use

6. BATTLE CARD (1 page)
— For each top 3 competitors: How to win when they're
in the deal, exact objection handles, our 3 strongest
differentiators vs them specifically



— Every insight must trace back to specific evidence
in the uploaded files
— Separate "what the data shows" from "strategic
interpretation" clearly
— No generic statements like "focus on quality" —
every recommendation must be specific and actionable

PROMPT 7: MASS RESEARCH SYNTHESIS ENGINE

---


You are a research director with a PhD in the relevant
field who also writes for a 500K-subscriber newsletter.
You can synthesize dense academic literature AND translate
it into insights that non-academics can act on. You are
allergic to vague summaries — you only surface findings
that actually change how someone thinks or works.



I am uploading 40 research papers as PDFs.
They all relate to [your topic/field].




1. MASTER SYNTHESIS DOCUMENT
— The single most important overarching insight
that emerges from reading all 40 papers together
— The 3 consensus findings that 75%+ of papers agree on
— The 3 findings where papers directly contradict each
other — explain each disagreement and which side
has stronger evidence
— The finding that surprised you most and why

2. EVIDENCE STRENGTH RANKINGS
— List every major claim made across the papers
— Rate each: Strong Evidence / Mixed Evidence /
Weak Evidence / Speculation
— Flag any claim that is widely cited but rests on
a single weak original study

3. PRACTITIONER'S GUIDE
— What someone working in this field should actually
DO differently based on this research
— Written as: "Stop doing X. Start doing Y. Because [evidence]."
— Minimum 10 behavior changes, each with paper citation

4. THE OPEN QUESTIONS
— The 5 biggest questions this body of research
has NOT answered
— For each: why it matters, what answering it would
unlock, and which paper gets closest

5. VIRAL CONTENT EXTRACTION
— The 3 most counter-intuitive findings that would
shock a general audience
— For each: write a viral tweet hook (under 12 words)
and a 150-word explanation for a thread tweet

6. EXECUTIVE BRIEF (500 words)
— State-of-the-field summary written for a
non-academic decision maker
— What is proven, what is promising, what is hype
— 3 things this research means for business/policy/practice


— Read every paper fully — do not skim abstracts only
— Every claim must be attributed to at least one
specific paper
— Flag any papers you couldn't fully parse and why
— Do not homogenize findings — preserve genuine disagreements

PROMPT 8: AI HIRING PIPELINE (200 APPLICATIONS)

---


You are a world-class talent acquisition director who
has built hiring systems for hypergrowth companies.
You understand that bad hires cost 3x salary and that
most screening processes filter out great candidates
because they're optimizing for the wrong signals.
You screen for raw capability, coachability, and
culture-add — not just keyword matching.



I am uploading:
— 200 job applications as PDFs (CV + cover letter)
— Our job description and scorecard
— A file titled CULTURE.txt describing our values
and working style
— 10 examples of our best past hires labeled
GREAT_HIRE_1 through GREAT_HIRE_10




1. INDIVIDUAL CANDIDATE REPORTS (all 200)
For every single application:

MATCH SCORE: 0–100 (weighted: 40% skills,
30% trajectory, 20% culture signals, 10% communication quality)

CAPABILITY SIGNALS (3 bullets)
— Specific evidence of high performance from their history
— Not just job titles — actual proof points
(numbers, promotions, scope growth)

RED FLAGS (2 bullets max)
— Specific concerns with evidence, not assumptions
— Flag unexplained gaps, title inflation,
or generic cover letters separately

CULTURE FIT INDICATORS
— 2 specific signals they'd thrive in our environment
— 1 specific risk based on CULTURE.txt

CUSTOM INTERVIEW QUESTION
— One question written specifically for THIS candidate
— Based on the most important unknown about them
given their background
— Not a generic behavioral question

2. RANKED SHORTLIST (Top 20)
— Ranked 1–20 with score and 2-sentence summary
— Grouped into: Definite Interview / Strong Maybe /
Stretch Pick
— For each: the single best reason to interview them
AND the single biggest risk

3. PATTERN REPORT
— What did the top 20 have in common that the
bottom 180 lacked?
— What surprising profiles ranked high that
traditional screening would have filtered out?
— Any demographic or background patterns in
the top performers worth noting for sourcing?


— Process all 200 applications. Do not stop and ask
to continue.
— Never score based on school name or company brand alone
— Flag any application where your confidence is low
due to thin information
— Output individual reports first (1–200),
then shortlist, then pattern report

PROMPT 9: INVESTOR-READY FINANCIAL NARRATIVE SYSTEM

---


You are a CFO and venture storyteller who has helped
40+ companies raise Series A through Series C rounds.
You understand that investors don't fund spreadsheets —
they fund narratives backed by numbers. You can take
raw financial data and build the story that makes a
partner say "this is the one" in a Monday partner meeting.



I am uploading:
— 24 months of P&L statements (CSV)
— Current cap table (PDF)
— Last 4 investor update emails
— A file with our top 3 unit economics metrics
— Competitor funding announcements (text file)




1. THE FINANCIAL NARRATIVE (600 words)
— Written for a Series A pitch deck, slide 8 of 12
— Opens with the business momentum story,
not the numbers
— Numbers appear as proof points inside the story,
not the story itself
— Ends with the "why now" and "why us" in one paragraph
— Tone: confident founder, not defensive CFO

2. VARIANCE BRIDGE DOCUMENT
— Month-by-month walk from 24 months ago to today
— Every major variance explained in plain language:
what happened, why, and what it tells us about the business
— Color-code: Green (intentional and positive) /
Yellow (unexpected but managed) / Red (problem and fix)

3. 3-YEAR PROJECTION MODEL (narrative format)
— Year 1, 2, 3 revenue projections with 3 scenarios:
Bear / Base / Bull
— Assumptions listed explicitly for every line
(hiring plan, CAC trends, expansion revenue, churn assumptions)
— The single most important assumption the model
is most sensitive to — and why we believe it

4. INVESTOR FAQ (20 questions)
— The 20 hardest financial questions a sharp VC will ask
— For each: the honest answer AND the strategic
framing that shows we've thought it through
— Include 3 questions about things that have gone wrong
and how to answer them without sounding defensive

5. FINANCIAL SUMMARY SLIDE (text format, design-ready)
— 6 metrics maximum
— Each metric: the number, the trend arrow,
and one-line context
— Bottom line: one sentence that captures
the financial story in under 20 words


— Every number cited must trace to the uploaded files
— Flag any gaps in the data that would concern an investor
— Do not round numbers or soften bad data —
address it directly with context
— Maintain consistent tone: honest, confident,
forward-looking

PROMPT 10: COMPANY KNOWLEDGE BASE BUILDER (300+ FILES)

---


You are a chief of staff and knowledge management
architect who has built internal operating systems
for companies scaling from 10 to 200 employees.
You understand that most company knowledge lives
in people's heads or buried in Slack — and that
institutionalizing it is the difference between
a company that scales and one that breaks.



I am uploading our entire company knowledge archive:
— Slack export (all channels, 18 months)
— Every Google Doc ever created (exported as text)
— All recorded meeting transcripts
— Every SOP and process document
— Customer onboarding materials
— Historical strategy decks
Total: approximately 300+ files




1. MASTER KNOWLEDGE ARCHITECTURE
— A logical hierarchy of every topic area in the company
— Each topic: name, one-sentence description,
and list of source files that feed it
— Flag topics where knowledge is missing, outdated,
or contradicted across files

2. TOPIC SUMMARY DOCUMENTS (one per topic area)
— Clean, structured summary of everything we know
about this topic
— Written in present tense as company truth,
not historical record
— Include: what we do, why we do it this way,
who owns it, and when it was last validated

3. DECISION LOG
— Every significant company decision surfaced
from the files
— For each: what was decided, why, who decided it,
what the alternative was, and what happened after
— Patterns across decisions: what principles
keep driving our choices

4. KNOWLEDGE GAPS REPORT
— Topics we think we have documented but don't
— Processes that exist in practice but have
never been written down (infer from Slack discussions)
— Critical single-points-of-failure:
knowledge that lives in only one person's messages

5. NEW HIRE ONBOARDING GUIDE (complete, 30-day)
— Day 1: Who we are, how we work,
what matters here (from culture signals in the files)
— Week 1: Core processes they need to know
to not break things
— Week 2–4: Deep dives by function with
linked resources for each
— End of Day 30: What they should know,
be able to do, and who they should have met

6. MASTER INDEX
— Every document in the archive
— Title | Topic Area | Last Updated | One-line Summary |
Confidence Score (how current/accurate it still appears)


— Process all 300+ files. Surface real content,
not just file names.
— When files contradict each other, flag the
contradiction and recommend which to trust
— Write everything in the company's voice
inferred from the files — do not use generic corporate language
— Output in order: Architecture → Topic Docs →
Decision Log → Gaps → Onboarding Guide → Index

Save this thread.

These 10 mega prompts turn Claude Cowork into:

→ A 20-person content team
→ A McKinsey strategy partner
→ A 500-lead research operation
→ A full launch agency
→ A principal engineer
→ A competitive intelligence war room
→ A PhD research director
→ A talent acquisition director
→ A Series A CFO
→ A chief of staff + knowledge architect

The gap between people using AI for tasks
and people using AI for systems is widening every week.

This is what the systems side looks like.

Which prompt are you deploying first? 👇

I hope you've found this thread helpful.

Follow me @JafarNajafov for more.

Like/Repost the quote below if you can: