And for videos it’s even worse because that’s a whole day gone, or you’re hiring someone to do it for you.

So I built something stupid simple: an image generator AND a video generator that both run inside Cursor (or Claude Code, whatever you’re using, same idea).

Taste was good but execution still needs some work. I went for that whole retro aesthetic that’s everywhere in developer tools right now, some people call it 90s vibe, some call it dithered, whatever it is it looks cool and makes your brand feel more relatable to devs.

I’m going to show you exactly how I built both, and it took about an hour total.

Why this aesthetic matters right now

The best developer tools are all using this terminal/ASCII aesthetic with dithering and film grain. skills.sh, Claude Code, tavus.io. It's relatable to devs who live inside CLIs all day. Your brand reflects your audience's actual environment.

skills.sh does Unicode block text. Claude Code does terminal-style graphics. tavus.io does dithered halftone with retro window frames.

Part 1: the image generator

What I actually built

The setup is super simple. I say “create a newsletter header for ABM WEEKLY with that retro dithered look” in Cursor. Python script generates Unicode block text, applies dithering, scanlines, grain, adds branding, outputs a 1200x630 PNG. No Figma, no Canva, no switching apps, I just stay in my editor.

The real numbers

Took about 30-45 minutes to get the whole thing working. Opus 4.5 wrote most of the 1200-line Python script, I added a Cursor skill so I could talk to it in English, and a config file with 13 color schemes. Now it's under 5 seconds per image.

How I actually built this

I didn't write a script. I had a conversation with Opus 4.5 and we built it iteratively.

I started with: “I want the aesthetic to be very similar to skills.sh. Something that Claude Code did, something that Cursor is also doing. It has to be super super cool.” And it built a first version, but it was missing something.

So I said: "You haven't added dithering effects or aesthetic. Like I also want a dithering type of pure aesthetic." And we added noise/grain texture for that film grain look, CRT scanlines for old monitor vibes, vignette for darkened corners and that cinematic feel, and dither textures like dots, ordered patterns, crosshatch.

And since I don't understand colors that well, I asked: "What is the hex code of that matrix green? Any more hex codes of colors that we might be using?" And we ended up building 13 different color schemes, from classic terminal to premium amber/gold to that warm tavus.io cream.

How it actually works

The Python script

I built newsletter_image_generator.py that handles everything. It does ASCII text rendering with Unicode block characters that look like the classic terminal art:

 █████╗ ██████╗ ███╗   ███╗
██╔══██╗██╔══██╗████╗ ████║
███████║██████╔╝██╔████╔██║
██╔══██║██╔══██╗██║╚██╔╝██║
██║  ██║██████╔╝██║ ╚═╝ ██║
╚═╝  ╚═╝╚═════╝ ╚═╝     ╚═╝

I went a bit overboard with 13 color schemes and dithering effects but whatever. And it automatically adds QBack logo and newsletter.qback.ai URL to every image so branding is handled.

So something like this.

The Cursor/Claude Code skill

I created a skill file so instead of CLI flags I just say "Create a newsletter header for AI OPS" or "Make it retro with that grainy film look." Skill figures out what flags to use.

The skill file (it looked like this)

Save this as .cursor/skills/newsletter-images/SKILL.md in your project. You’ll need the Python script and pip install Pillow pyfiglet first.

---
name: newsletter-images
description: Generate ASCII/terminal-style newsletter images with dithering effects. Use when user asks to create images, headers, graphics, or visuals for newsletters, LinkedIn, Twitter, or other social media. Also use when user mentions retro, dithered, grainy, vintage, scanlines, or CRT aesthetic.
---

# Newsletter image generator skill

When the user asks you to create, generate, or make images for newsletters or social media, use this skill.

## Color selection workflow

The user is not a designer. Ask about the vibe first:

> "What vibe are you going for?
> - classic hacker/builder (black background, white/green text)
> - matrix/AI energy (black with neon green)
> - warm & human (cream background with coral accent)
> - premium/luxury (black with amber/gold text)
>
> Want me to generate quick samples?"

## Quick reference

| User says | Use this |
|---|---|
| "hacker", "builder", "classic" | `--style dark_terminal` |
| "matrix", "AI", "green" | `--style dark_matrix` |
| "warm", "human", "friendly" | `--style tavus_warm` |
| "premium", "luxury", "gold" | `--style amber_noir` |
| "retro", "vintage" | `--retro` |
| "grainy", "film grain" | `--noise` |
| "scanlines", "CRT" | `--scanlines` |

## Basic command

```bash
python newsletter_image_generator.py --title "TITLE" --ascii-font unicode_block [options]
```

## Examples

```bash
# Newsletter header
python newsletter_image_generator.py --title "ABM WEEKLY" --ascii-font unicode_block --style dark_terminal

# Full retro
python newsletter_image_generator.py --title "AI OPS" --ascii-font unicode_block --retro

# Matrix + retro + dots
python newsletter_image_generator.py --title "DITHER" --ascii-font unicode_block --style dark_matrix --retro --texture dots
```

PART 2: Video Generation

You might've seen the Remotion Skills video that went viral on X. I built a video generator inside Cursor using the same approach.

The prompts I used with Opus 4.5 (thinking)

The initial prompt was: “Create a Remotion animation for a newsletter announcement. Frame the entire content inside a stylized email client mockup (Gmail-style). Show an email from ‘QBack Newsletter’ with subject line, then animate screenshots inside the email body area.”

Here is the GitHub gist to see the entire prompts + setup and my chat messages with Opus 4.5.

For the color scheme I specified: Background #FBF7F4 (Warm Cream), Text #2D2A26 (Charcoal), Accent #E85D4C (Coral Red), Frame/Border #E8E0D8 (Warm Gray).

For the animation style I said: “Make it scroll-stopping with a notification badge that bounces, a like counter that increments, progress dots for slides, cards that slide up from below with slight rotation, exit by sliding up and fading out, and a subtle breathing animation on the window.”

Video Setup

1. Create Remotion project

npx create-video@latest
cd your-project

2. Add your assets

Place in /public folder: newsletter-header.png for your newsletter header image, then screenshot-1.png, screenshot-2.png, and screenshot-3.png for your content screenshots.

3. The core components

The video is built with animated React components.
GrainOverlay handles the film grain texture using an SVG filter for fractal noise that changes every few frames.
NotificationBadge creates that bouncing notification on the sender avatar using spring animations.
LikeCounter is an animated heart that fills and increments as likes go up. ProgressDots shows TikTok-style slide indicators where the active dot expands. And EmailClientFrame is the Gmail-style wrapper with traffic lights, sidebar, and everything.

const GrainOverlay: React.FC<{ frame: number; intensity?: number }> = ({ frame, intensity = 0.05 }) => {
  const seed = Math.floor(frame / 3);
  // SVG filter for fractal noise that changes every few frames
};

const NotificationBadge: React.FC<{ frame: number; fps: number }> = ({ frame, fps }) => {
  const badgeIn = spring({ frame: Math.max(0, frame - 5), fps, config: { damping: 10, stiffness: 200 } });
  const bounce = Math.sin(frame * 0.15) * 2;
  // Returns a bouncing "1" badge
};

const LikeCounter: React.FC<{ frame: number; fps: number; startFrame: number }> = ({ frame, fps, startFrame }) => {
  const count = Math.min(Math.floor(progress / 3), 7);
  // Heart changes from 🤍 to ❤️ as likes increase
};

4. Register and Render

// Root.tsx
export const RemotionRoot: React.FC = () => {
  return (
    <Composition
      id="NewsletterAnnouncement"
      component={QBackNewsletterAnnouncement}
      durationInFrames={360}
      fps={30}
      width={1280}
      height={720}
    />
  );
};

npx remotion render NewsletterAnnouncement out/newsletter.mp4

What this CAN’T do

Let me be real about the limitations because there are definitely things this setup can’t handle.

• You can’t do 3D rendering. You can’t create images like Clicker’s 3D renders because that would need actual 3D software or an image LLM where you keep prompting until you get what you want, which is a different workflow entirely.

• You also can’t do complex photo manipulation since this is ASCII/terminal aesthetic only. Photorealistic stuff is Midjourney or DALL-E territory.

• It’s also simple compositions only, meaning one title, optional subtitle, optional window frame, not a full design tool. And videos need assets so you still need to create the screenshots that go inside the frames.

But for newsletter headers, social media images, and announcement videos in the developer/builder aesthetic, it’s exactly what you need.

The stack

For images you need Python 3.x, Pillow for image manipulation (pip install Pillow), and pyfiglet for ASCII art generation (pip install pyfiglet). For videos you need Node.js, Remotion which is a React-based video framework (npx create-video@latest), and React which comes with Remotion.

That’s it. No API keys, no subscriptions, no cloud services.

Who should use this?

Dev tools companies probably already doing something similar (for images). Newsletter writers who want consistent headers without design skills. Anyone tired of Canva. Anyone who wants the same colors and format every time. Or wants to create instant product videos.

Happy building, signing off your quarterback! 🏈

And I finally cracked it.

Here’s how the growth visibility looks like now:

For months (thanks to my team, Ishika and Fredson) I’ve been testing AEO, SEO, GEO, AI content strategies – basically everything to become the “source of truth” that ChatGPT, Claude, Perplexity actually cite.

Most marketers are trying to game AI search. I built a system that makes AI models cite you naturally.

But first, let’s be clear about WHAT content to optimize for LLMs.

Two types of content: Company blog vs LLM-indexable content

Your company blog (authentic, honest, and product-specific content):

This is where you share authentic company stories that people actually want to read:

• Fundraise announcements and company milestones

• Product launches and feature releases

• Customer case studies (”How X achieved 3x conversion using [your product]”)

• Behind-the-scenes content (how you shipped a launch, what went wrong, what worked)

• Engineering learnings and technical deep-dives

• Marketing experiments and results

• Company culture and team stories

This content lives on your /blog, /customers, /team, /engineering, /inside and gets shared on social. It builds brand, drives engagement, and creates authentic connections. Best example for top tier content on this is posthog.

LLM-indexable content (programmatic SEO at scale):

This is where you become the source of truth for AI search engines:

• Tool-related pages: Integration guides, comparison pages, “how to use X with Y”

• Pain point solutions: Solving specific problems your buyers search for

• Template libraries: Resources, calculators, generators your buyers need

• Information hubs: Comprehensive guides answering buyer questions at scale

This content is:

• Closely connected to your product (every page leads to using your product)

• Scalable (you can create hundreds/thousands of variations)

• Built programmatically (using Next.js, templates, automation)

• Optimized for LLM parsing (FAQ schema, structured data, comparative analysis)

The key difference: Your company blog is about YOU. LLM-indexable content is about SOLVING PROBLEMS your buyers search for.

If you’re trying to rank on ChatGPT/Claude/Perplexity, focus on programmatic SEO content that’s specific to your product and scalable. Think Zapier’s integration pages, not their company blog.

The journey to solve this began with 👇 this Lovable app that I built 4 months ago.

What’s working at scale (the proof)

Here’s what makes these companies successful - every single page connects directly to their core product and solves a real customer problem.

This isn’t random content at scale. It’s solving problems your customers actually have, using your product.

Zapier’s playbook: They created integration pages for specific app pairings. Pages like “Connect Slack to Google Sheets”, “Trello to Gmail integration”, “Airtable to HubSpot” – all programmatically generated using their integration data.

URL pattern: zapier.com/apps/{app1}/integrations/{app2}

In fact, I got inspired by them to create X to Y for an ETL startup I was working in 2018. I created the entire playbook to make sure if anyone searches for terms like Postgres to Redshift migration then this page would 1000% show up on top 10 on google or chatgpt for sure. It is still the money maker today.

Vendr.com’s strategy: They built pricing pages for every SaaS tool. As everyone knows it’s super hard to find pricing for companies which hide it behind “talk to sales”. “Tegus pricing”, “Radar Labs pricing” – hundreds of pages that answer the exact questions buyers search for when comparing tools.

URL pattern: vendr.com/marketplace/{tool-name}

Clay’s approach: Enrichment playbooks and data workflows at scale. Clay creates pages for specific enrichment use cases and data operations that connect directly to their product – targeting sales and marketing teams searching for data enrichment solutions.

The questions these pages answer come from:

• Customer support tickets (”How do I connect Slack to Google Sheets?”)

• Sales calls (”Do you have a template for admin dashboards?”)

• G2/Reddit threads (”What’s the best way to convert JSON to CSV?”)

• Your actual product usage data (what features/integrations customers search for most)

If your programmatic content doesn’t connect to your product AND solve a verified customer problem, you’re just creating SEO spam. LLMs won’t trust it, and users won’t find it valuable.

The exact process to create this yourself

Here’s the 5-step system that makes AI models cite you:

Step 1: Gather buyer & competitor intelligence

Before creating any content, you need to understand what questions your buyers are actually asking and where competitors are failing to answer them.

You need three data sources:

Data source #1: Your buyer intelligence (MOST IMPORTANT)

This is where you find questions that connect to your product:

• Sales call recordings (Gong, Zoom data) - “How do I...?” questions

• Support conversations (Intercom, Zendesk) - Problems customers are trying to solve WITH your product

• Customer interviews and feedback - What they wish your product could help them do

• Product usage data - What features/use cases they search for most

• Onboarding questions - What new users struggle to understand

Critical: Every question here should have an answer that involves using your product. If customers are asking about it, and your product solves it, that’s programmatic content gold.

Data source #2: Community intelligence (where your buyers hang out)

Mine for problems your product solves:

• Reddit discussions in your category - What are people asking about?

• G2/Gartner reviews (yours + competitors) - What features do people wish existed? (i know they don’t matter in today’s world but still you need it for data)

• LinkedIn conversations - What problems are people talking about in your space?

• Twitter threads about your space - What frustrations come up repeatedly?

Critical: Filter for problems YOUR product actually solves. Don’t create content about problems unrelated to your product.

Data source #3: Competitor intelligence (find the gaps)

Track:

• What pages are they creating? (Do they connect to their product?)

• What questions are they answering? (Can you answer better with your data?)

• Where are the gaps in their content? (Problems they’re not addressing that your product solves)

The filter: Every content idea must pass this test:

1. Is this a real question customers/prospects ask? (Verified in data)

2. Does our product solve this problem? (Direct connection)

3. Can we be the authoritative source? (We have unique data/experience)

4. Will LLMs trust this as factual? (Data-backed, not opinion)

PROMPT #1: Extract content opportunities from buyer intelligence

Here’s the exact prompt I use to extract patterns from customer calls:

You are an elite B2B SaaS content strategist and competitive intelligence analyst. You specialize in transforming buyer intelligence data into high-converting programmatic content strategies. You have deep expertise in:
- LLM/AEO optimization (making content rank on ChatGPT, Claude, Perplexity)
- Competitive gap analysis and differentiation strategy
- Bottom-funnel content that directly influences purchase decisions
- Scalable content architecture (templates, pillars, automation)

Your task: Analyze all sales call transcripts and generate a complete programmatic content roadmap with competitive positioning.

Use the data from

Pain Points Identified
Sales Objections
Use Cases
Benefits mentioned
Persona
Competitor comparisons


---

ANALYSIS FRAMEWORK:

**Phase 1: Content Opportunity Extraction**

For each pain point, objection, use case, and comparison:
1. Convert to buyer-intent search queries (both Google and LLM format)
   - Google: “how to solve [pain point]”
   - LLM: “can you help me with [pain point], I’m using [current solution]”
2. Identify the natural question format buyers ask
3. Map to buyer journey stage (awareness/consideration/decision)
4. Score LLM ranking potential (1-10)

**Phase 2: Competitive Gap Analysis**

For each identified content opportunity:
1. Research what [COMPETITOR 1] currently publish
2. Rate their content quality (1-10) based on:
   - Depth of answer
   - Use of specific data vs generic advice
   - LLM optimization (FAQ format, structured data)
   - Recency and accuracy
3. Identify specific gaps:
   - What questions do they NOT answer?
   - Where do they use generic advice vs specific data?
   - What objections/comparisons do they avoid addressing?
4. Determine our differentiation angle (based on our unique data)

**Phase 3: Scalability Assessment**

For each content opportunity:
1. Can this be templated? (Yes/No + reasoning)
2. How many variations can we create? (Calculate: pain point × persona × use case)
3. What variables enable programmatic scaling?
4. What unique data makes each variation valuable?

**Phase 4: Comparative Content Strategy**

Create head-to-head content recommendations:
1. “Us vs [Competitor]” comparison pages
2. “[Competitor] alternative” positioning pages
3. “Best [category] for [use case]” roundup pages (featuring us prominently)
4. Objection-handling pages that address competitive concerns

---

OUTPUT FORMAT:

**Part 1: Content Opportunity Matrix**

Create a table with these columns:
| Content Opportunity (Search Query) | Pain Point/Objection Addressed | LLM Query Format | Our Differentiation Angle| Priority Score |

**Part 2: Top 30 High-Priority Content Opportunities** (keep it crisp)

For each, provide:

---
**Opportunity #[X]: [Content Title]**

**Search Intent:** [What buyer is trying to accomplish]
**Triggered By:** [Which pain point/objection/use case from your data]
**Target Persona:** [Specific role + context from use cases]
**Buyer Journey Stage:** [Awareness/Consideration/Decision]

**Competitive Analysis:**
- [Competitor 1]: [Current content + quality score + specific gap]
- **Gap Opportunity:** [What we can create that they can’t/won’t]

**Our Unique Data Advantage:**
[Specific proprietary data/insights from our calls that competitors lack]

---

**Part 3: Comparative Content Strategy**

Generate 10 “Us vs Competitor” content opportunities:

| Page Title | Competitor Targeted | Key Differentiator (from our data) | Objection Addressed

**Part 4: Quick Wins (Immediate Implementation)**

List 10 content pieces we can create THIS WEEK:
1. [Title] - Uses [specific data point] - Addresses [competitor gap] - Effort: [X hours]
2. [...]

---

CONSTRAINTS:
- Only suggest content where we have unique proprietary data
- Every recommendation must address a specific competitor gap
- All content must be optimized for LLM ranking (FAQ format, structured data)
- Focus on bottom-funnel, high-conversion content
- Must be scalable to 20+ page variations minimum

SUCCESS CRITERIA:
Your output enables us to:
1. Create content that competitors CAN’T replicate (due to our unique data)
2. Directly address buyer objections/comparisons with data-backed answers
3. Rank on ChatGPT/Claude/Perplexity for high-intent queries
4. Scale to 10000+ pages within 30 days using templates
5. Differentiate from competitors with specific, factual content

This gives you a content roadmap based on what people actually care about – not what you think they should care about.

Step 2: Extract content opportunities & pillars

Now that you have your buyer intelligence, use it to identify what content categories (pillars) you can create at scale.

This isn’t about tricking algorithms. It’s about being THE authoritative source in your space so AI has no choice but to cite you.

PROMPT #2: Identify programmatic content categories from content opportunities

Use this prompt to transform your content opportunities (from PROMPT #1) into scalable content pillars:

You are an elite programmatic SEO architect and LLM optimization specialist. You specialize in creating scalable content systems that rank on AI search engines (ChatGPT, Claude, Perplexity, Gemini). Your expertise includes:
- Content template design and variable identification for programmatic scaling
- LLM/AEO optimization (FAQ schemas, structured data, natural language queries)
- Bottom-funnel content strategy that drives conversions
- Competitive moat-building through proprietary data-driven content

Your task: Transform identified content opportunities into 15 scalable content categories (pillars) with complete template specifications and automation potential.

---

Refer to the file
---

ANALYSIS FRAMEWORK:

**Phase 1: Pattern Recognition & Clustering**

Group similar content opportunities into categories based on:
1. **Content type pattern** (e.g., “How to integrate X with Y”, “[Competitor] alternative for [use case]”, “[Use case] template”)
2. **Scalability variables** (what can we systematically vary? Industry, company size, integration, use case, persona)
3. **Product connection strength** (does our product naturally solve this? Can we demonstrate unique value?)
4. **LLM ranking potential** (FAQ-friendly? Data-rich? Comparative?)

**Phase 2: Template Architecture Design**

For each content category, define:
1. **URL structure** with variables: `/[category]/[variable-1]/[variable-2]`
   - Example: `/integrations/{tool-1}/{tool-2}` (Zapier model)
   - Example: `/marketplace/{vendor}/pricing` (Vendr model)
   - Example: `/templates/{use-case}/{industry}` (Canva model)

2. **Content template H2/H3 structure** optimized for LLM parsing:
   - Must include FAQ section (5-7 questions)




---

OUTPUT FORMAT:

**Part 1: Content Category Overview**

| Category Name | URL Pattern | Total Pages (calculation) | LLM Score | Competitive Diff Score 
**Part 2: Detailed Category Specifications (Top 5)**

For each category, provide:

---
**CATEGORY #[X]: [Category Name]**

**Content Type Pattern:** [Describe the pattern - e.g., “Integration guides”, “Competitor alternatives”, “Use case templates”]

**URL Structure:** `yourdomain.com/[category]/{variable-1}/{variable-2}/{variable-3}`

**Example URLs:**
1. [Specific example URL 1]
2. [Specific example URL 2]
3. [Specific example URL 3]


**Content Opportunities Addressed:
- use specific opportunities from the file that this category covers]

**Unique Data Requirements:**
- Data source 1: [What proprietary data is needed - e.g., “Customer success metrics by integration”]
- Data source 2: [e.g., “Competitive pricing analysis from 500+ customers”]
- Data source 3: [e.g., “Time-to-value benchmarks by use case”]


**Content Template Structure:**

```markdown
URL: /[category]/{variable-1}/{variable-2}

CONSTRAINTS:

Only suggest categories where we have proprietary data competitors can’t easily replicate

Every category must scale to 50+ page variations minimum

Every category must connect directly to our product (natural CTA/conversion path)

Focus on bottom-funnel, high-intent content types

All templates must be optimized for LLM parsing (FAQ schema, structured data)

SUCCESS CRITERIA: Your output enables us to:

Build 500+ pages in 30 days using programmatic templates

Rank on ChatGPT/Claude/Perplexity for thousands of long-tail queries

Create content with lasting competitive differentiation (based on unique data)

Automate 80%+ of content generation while maintaining quality

Directly influence purchase decisions with bottom-funnel content

Step 3: Build your content template system (AI writer + AI judge)

Once you have your content categories identified, you need two AI systems working together:

1. The Writer: Generates content at scale using your templates.

2. The Judge: Audits content for authenticity and accuracy

This prompt is for directional help as the prompt will vary a lot depending on industry and product.

PROMPT #3: The writer system (generate LLM-optimized content at scale)


You are an elite B2B SaaS content writer and LLM optimization specialist. You specialize in creating data-driven, authoritative content that ranks on AI search engines (ChatGPT, Claude, Perplexity, Gemini). Your expertise includes:

Your task: Generate a complete content page following the provided template, using proprietary data to create content competitors cannot replicate.

Writing FAQ-optimized content that LLMs cite as authoritative sources

Comparative analysis and competitive positioning in content

Data storytelling with proprietary metrics and customer insights

Bottom-funnel content that directly influences purchase decisions

Technical accuracy and fact-checking for B2B audiences


Voice & Tone:

Authoritative but conversational (write like you’re explaining to a smart colleague)

Use “we analyzed X companies” not “companies typically”

Use “you” to address reader directly

NO marketing fluff or sales language

NO adjectives without data (don’t say “powerful”, “easy”, “revolutionary”)

Data Requirements (MANDATORY):

Every claim MUST cite specific source

Use exact numbers: “Based on 247 customer calls” not “hundreds of calls”

Include your proprietary data prominently (”Our analysis of...”)

Acknowledge data limitations honestly when appropriate

Link to external sources for additional credibility

LLM/AEO Optimization (CRITICAL):

PRIMARY keyword: Include in H1, first 100 words, naturally throughout (3-5 times total)

SECONDARY keywords: Include in H2 headers and body (1-2 times each)

LLM QUERY PHRASES: Integrate naturally (10-15 phrase variations)

“how do I [action]”

“what’s the best way to [solve problem]”

“can you help me [accomplish goal]”

“how to [do thing] for [use case]”

Use question format for H2 headers where natural

Include comparative language (”vs”, “compared to”, “alternative to”)

FAQ Requirements (MANDATORY):

5-7 questions REQUIRED on every page

Use exact phrasing from buyer intelligence OR common LLM queries

Each answer 2-3 sentences with specific data

Include long-tail keyword variations in questions

Content Authenticity:

FORBIDDEN PHRASES: “In today’s digital landscape”, “leverage”, “synergy”, “game-changer”, “revolutionary”, “seamless”, “empower”

NO claims without data support

Mention your product ONCE naturally (don’t force it in every section)

Content must be specific to this topic (not generic advice)

All claims must be falsifiable (can be proven wrong if incorrect)

Phase 3: Self-Quality Check (Before Submitting)

Verify checklist: 
□ Every number is cited with specific source 
□ Zero generic AI-generated phrases 
□ Title matches how target audience actually searches 
□ First H2 provides immediate, data-backed answer 
□ FAQ includes 5-7 questions from buyer intelligence 
□ Content demonstrates unique expertise/data 
□ Zero promotional language 
□ YAML frontmatter complete 
□ Primary keyword density: 3-5 mentions (not keyword stuffed) 
□ Secondary keywords in H2 headers 
□ 10-15 natural language phrase variations included

CONSTRAINTS:

Only use data from sources provided (do NOT make up statistics)

If data is missing for a section, note it and suggest where to find it

Do not write promotional content disguised as advice

Must demonstrate unique value from proprietary data

Content must be specific enough that competitors can’t easily replicate

PROMPT #4: The judge system (quality audit at scale)

This is the key to maintaining quality at scale. nailed it recently: “AI has flooded channels with lookalike content. If you can’t make something that only your company can make, don’t make it.”

That’s exactly what the Judge system prevents. After content generation, run it through this comprehensive audit

You are an elite B2B SaaS content auditor, fact-checker, and LLM optimization analyst. You specialize in quality assurance for programmatic content at scale. Your expertise includes:

Fact-checking and data verification for technical B2B content

Detecting AI-generated generic language and maintaining authentic voice

LLM/AEO optimization analysis (FAQPage schema, featured snippet readiness)

Competitive differentiation assessment (unique data validation)

Bottom-funnel content performance optimization

Your task: Audit generated content against source materials and provide a PASS/REVISE/REJECT decision with specific revision instructions.

AUDIT FRAMEWORK (Score Each Section 1-10):

FACTUAL ACCURACY AUDIT For every claim, data point, and statistic: □ Verify it exists in source materials (quote exact source location) □ Check if numbers are calculated correctly □ Confirm dates and timeframes are accurate □ Validate that quotes are exact (not paraphrased incorrectly) □ Flag any claims made without data support □ Cross-reference competitor information cited □ Check for outdated information (flag anything >6 months old)

List ALL factual issues found with:

Location: [Section/Paragraph]

Issue: [What’s wrong]

Source Check: [Verified in source? Yes/No/Partially]

Correction: [Exact corrected text]

Severity: [Critical/Moderate/Minor]

AUTHENTICITY AUDIT Scan for generic AI-generated language: □ “In today’s digital landscape” or similar filler phrases □ Buzzwords without substance (leverage, synergy, game-changer) □ Vague statements: “many companies”, “experts say”, “it’s important to” □ Claims that could apply to any company in any industry □ Overly smooth transitions that feel artificial □ Perfect grammar that lacks human voice □ Lists without supporting evidence

List ALL authenticity issues:

Generic phrase: “[EXACT TEXT]”

Why it’s generic: [Explanation]

Rewrite with specific data: “[NEW VERSION WITH NUMBERS/NAMES/SPECIFICS]”

COMPLETENESS AUDIT Check against customer questions/source data: □ Are all key customer concerns addressed? □ Are there obvious follow-up questions not answered? □ Is competitive context missing? □ Are implementation details vague or missing? □ Are there data gaps where we should have information? □ Are edge cases or limitations acknowledged?

List missing elements:

Missing: [What’s not covered]

Why it matters: [Impact on usefulness]

Source available: [Yes/No - where to find this information]

Suggested addition: [2-3 sentence summary of what to add]

LLM OPTIMIZATION AUDIT Evaluate for LLM ranking potential: □ Are headers phrased as questions people actually ask? □ Is the immediate answer provided in first 100 words? □ Is content structured with clear hierarchy (H1>H2>H3)? □ Are there specific data points LLMs can extract and cite? □ Is there a FAQ section with exact customer questions? □ Are there citations to authoritative external sources? □ Is the content factual vs opinion-based? □ Can key information be pulled for featured snippets?

LLM Ranking Score: [1-10] Breakdown:

Question format: [Score + explanation]

Immediate answer: [Score + explanation]

Data density: [Score + explanation]

Citation quality: [Score + explanation]

FAQ strength: [Score + explanation]

PROMOTIONAL LANGUAGE AUDIT Check for sales pitch disguised as content: □ Is our product mentioned more than once? □ Is our product mentioned before addressing the core question? □ Are benefits described without data support? □ Is competitive comparison unfairly biased? □ Are there CTAs before value is delivered? □ Is the content useful even to non-customers?

List promotional issues:

Location: [Section]

Promotional language: “[EXACT TEXT]”

Why it’s promotional: [Explanation]

Neutral rewrite: “[VERSION WITHOUT PROMOTION]”

This two-system approach lets you scale content while maintaining the quality that makes you a trusted source for LLMs.

Step 4: Technical SEO/AEO/GEO optimization (the indexing hacks)

Creating great content isn’t enough. You need to make sure LLMs can actually find, parse, and cite your content.

Here are the technical hacks that get your pages indexed on ChatGPT, Claude, and Perplexity faster:

4.1: Schema markup & structured data

LLMs love structured data they can parse. Add JSON-LD schema to every page:

FAQPage schema (required on EVERY page):

Here is the prompt for GEO optimization check:

TASK: You are reviewing content for optimization in Generative Engine search results (e.g., Perplexity.ai, BingChat), using the COSTAR framework:
COSTAR =
C - Citations & Quotes:
Add inline citations or credible quotes relevant to the content to enhance reliability.Attribute every significant claim or statistic with a credible source.O - Optimization (Fluency and GEO-specific optimization):
Rewrite content to naturally incorporate provided keywords and closely match the intent of generative engine queries.Avoid unnatural keyword stuffing; instead, focus on smooth integration of keywords.S - Statistics:
Include quantitative data and statistics wherever possible, replacing general statements with specific figures to increase credibility.
T - Technical Terms:
Integrate relevant, domain-specific technical terms to convey expertise, aligning closely with the provided keywords.
A - Authoritative Language:
Edit content tone to appear authoritative and confident, highlighting unique insights clearly and persuasively without excessive exaggeration.
R - Readability:
Ensure the content is easy to read, structured clearly, and optimized for skimming.Use short paragraphs, subheadings, and bullet points effectively to aid readability.
HOW TO RESPOND:
Do not rewrite any sentences directly.Do not change or reword the original content.Instead, provide GEO optimization suggestions inline using this format:[Text block you’re commenting on]
{Your optimization suggestion or explanation here}
If you’re recommending a word or phrase change, show the change clearly:[helping you get insights quicker]
{data analytics instead of insights}
If suggesting a replacement for a heading or title:[OpenAI, But Make it Blazing Fast: Oracle + OpenAI is Here]
{replace it with: OpenAI for Oracle: 33% Increase in Performance and Optimized Capacity}
Only return inline suggestions across the original text.Avoid summary responses. Be specific and structural.
WHEN READY: Once the user shares:
Primary keywords (2-3)Secondary keywords (3-5)Original blog titleFull blog contentBegin providing inline COSTAR suggestions using the format above.

Output Goal: A marked-up version of the original content with precise GEO recommendations for ranking, clarity, and performance in generative search engines.

4.2: LLM-specific files (CRITICAL)

These files tell LLM crawlers what content to index and how to access it.

llms.txt file setup:

Create /llms.txt in your root directory:

YAML with all metadata:

Use the following prompt to update YAML content.

Act like the LLM specialist and browse my blog URL <insert URL> and fill up the following and tell me steps to where to paste this for YAML so that When a generative-engine crawler (or any system that later trains / retrieves with an LLM) meets that top-of-file YAML block. Pick what fits depending on the URL and page contents. Give me output in comma separated.

robots.txt Optimization:

Update your `/robots.txt` to allow LLM crawlers. Double-check if you run on Cloudflare as they blocked some really good LLM crawlers and I had to unblock it.

Phrase optimization (natural language queries):

Include variations of how people ask questions in LLMs.

Q&A format structure:

Structure content in question-answer pairs:

4.3: Infrastructure & fast indexing hacks

GitHub Pages indexing hack (for developer marketing)

If you’re targeting developers, GitHub pages get indexed by LLMs significantly faster:

1. Create a GitHub repository for your content

2. Enable GitHub Pages in repo settings

3. Structure as documentation

4. Add technical content (code samples, API docs, integration guides, benchmarks)

Why it works: LLMs prioritize developer documentation from GitHub because it’s typically high-quality, well-structured, and regularly updated.

Documentation pages get scraped faster by LLMs. Convert your docs to a dedicated `llms.txt` format:

Complete technical optimization checklist:

For each piece of content:

• ✅ YAML frontmatter with all metadata

• ✅ JSON-LD schema (FAQPage + Article/HowTo)

• ✅ Primary keyword in H1, first 100 words, naturally throughout

• ✅ Secondary keywords in H2 headers

• ✅ 10-15 natural language phrase variations for LLM queries

• ✅ Auto-generated FAQ section (5-7 questions)

• ✅ Meta description with primary keyword + data point

• ✅ OpenGraph and Twitter Card tags

• ✅ robots.txt allows all LLM crawlers

• ✅ llms.txt file includes page URL

• ✅ sitemap.xml updated automatically

• ✅ “Last Updated” timestamp in ISO 8601 format

• ✅ Author expertise statement

• ✅ Canonical URL specified

• ✅ Structured with clear H1 > H2 > H3 hierarchy

Step 5: Scale the system inside Cursor /Claude Code/ QBack (the full setup)

Here’s how to actually implement this inside Cursor (earlier) and QBack.

Forward this to your growth and product marketing friends who are still writing one blog post per week.

The marketers who figure out programmatic content + LLM ranking in 2026 will have an insane advantage over those who wait.

P.S. The technical optimization section (Step 4) is super important. The llms.txt file, robots.txt configuration, and JSON-LD schema are what actually get you indexed. I’ve seen companies create amazing content that never gets cited by LLMs because they skip the technical setup. Don’t skip it.

P.P.S. I also tested the recent bunch of GEO tools that launched. Some of them are straight-up gaslighting – charging premium prices for basic keyword stuffing wrapped in “AI optimization” language. The tools that actually work? The ones that help you structure your existing authoritative content for LLM parsing, not the ones that try to game citations with synthetic content.

Got your back, builders. 🎯

Want more content like this? Follow me on LinkedIn @harshitjain7 / Twitter @harshitqback where I share what’s actually working in AI marketing.

just called GTM in 2025 “a bit of a 🗑️🔥” - and he’s right. The tried-and-true channels don’t work like before. is coining new terms like “Gen Marketers” for generalists who can orchestrate high-impact campaigns. And Kyle Coleman’s sharing Forrester data showing companies saving 92,400 hours (equivalent to 46 full-time employees) through proper AI workspace convergence.

But here’s what they’re all dancing around without saying directly: the fundamental restructuring isn’t just about AI tools or new GTM motions. It’s about marketers who can build vs marketers who can’t.

The way I see it - how do you essentially become a future-proof product marketer? How do you not just survive but dominate in this new world?

Here’s the thing most people don’t get: we’re not in one market anymore. We’re in two completely different universes, and each requires totally different skills.

Universe 1: Pre-2023 tools scrambling to add AI

These are your Notions, Intercom, Salesforce - basically every SaaS tool you used in 2022. They’re all freaking out, adding AI layers on top of existing products, acquiring AI companies, trying not to get disrupted.

The positioning nightmare: What do you do when your AI features might eat your core product? How do you price it? How do you message the evolution without confusing existing buyers who just learned your old positioning?

Examples of companies doing this pivot:

• Notion repositioning around AI workspace intelligence

• Salesforce is doing acquisitions and building AI story on agentforce.

• Intercom evolving from chat to Fin AI

Where most PMMs fail: They don’t understand how AI actually works under the hood. So they can’t differentiate their AI implementation from competitors’. They end up with generic “AI-powered” messaging that means nothing.

What you need to master:

• How to reposition without alienating existing buyers

• Technical understanding of your AI implementation vs competitors

• Pricing strategy when AI features might cannibalize core product revenue

Universe 2: Post-2023 AI-native tools built different

These are your Cursors, Clays, Replit - tools built AI-first. They’re not adding AI features. AI IS the product.

The technical depth requirement: If you’re marketing these tools and you can’t explain the technical differentiators, you literally cannot sell them. Period.

For dev tools, you need to go deep: example: If you’re at Supabase - how do people actually use it? What’s the database performance vs Firebase? How does authentication work differently?

Otherwise, you will say “unclear or less informed” stuff like these when you don’t the deep tech and haven’t had a Github profile yet. Know the ICP and the product. 👇

For AI tools, you need to understand product and engineering:

• What “agentic” actually means (spoiler: it’s not just a chatbot)

• Model layer differences and why they matter

• Memory architecture and how it affects performance

• Latency management in reasoning loops

• Eval frameworks for systems that give different outputs each time

My simple product understanding workflow:

1. Use the product daily - build something real with it
2. Schedule engineering and product calls with 5-6 technical questions
3. Analyze Reddit/buyer calls for actual pain points
4. Run queries/workflows yourself to feel buyer pain
5. Check product analytics tool to see the user behavior

The 6 shifts that will define marketers in 2026

Based on what I’m seeing across both universes, here are the 7 fundamental shifts that will separate the marketers who thrive from those who get replaced:

1. The death of demand generation (and what replaces it)

Here’s my controversial take: Demand gen as a role is dead.

Why? Because I can literally replace most demand gen work (grunt work and I have done it too: ) with Clay and some custom automation.

What demand gen used to do:

• List building → Clay does this 100x faster

• Email sequences → AI writes better copy using buyer call data

• Lead scoring → Automated based on actual product usage data

• ABM campaigns → Clay + custom workflows with Webflow handle this at scale

What’s actually valuable now:

• Growth hacking - viral mechanics, creative experiments, influencer partnerships

• Creative content - video, explainer demos for each usecase, gamification

• Partnership playbook - community building, event experiences

• Performance optimization - ads, conversion rates, funnel analysis

The evolution: The future marketing org has two core roles:

Good growth hack example of what AI-first stack is capable of:

Performance marketing at scale: Fyxer went from $1M to $17M ARR in 8 months using this exact playbook (source: ):

• 200+ ads live at any given moment

• 100-150 new creative variations every two-week sprint

• Powered by AI tools: ChatGPT, Claude, and custom automation

• Key metric: cost-per-work-email (10x more valuable than personal emails)

  🏈 I share weekly marketing playbooks and AI apps for B2B product marketers. Don’t miss it!

Creative growth tactics (some examples) that actually work:

• Firecrawl gamifying onboarding with credits (get credits for completing steps)

• Dev tools creating toy store with messaging that resonates with buyers (wiz)

• Companies sending branded doormats for ABM conversation starters (delve) and their page was built by lovable.

Tcosystem opportunity: mentioned in her “Ecosystem Marketing” drive majority growth at startups - using 3rd parties like influencers, communities, channel partners, affiliates, integration partners, and buyers to piggyback off their reach and credibility.

The mindset shift: Think about how you can make your buyers feel like stars and show their work publicly. The idea of any tool is to help humans level up their lives - get better results, progress in their career, learn new skills.

As long as your buyers are having a great life because of your tool, you win.

Pick any of these 3 or hit me up on Twitter/LinkedIn with what you want next.

2. Speed is the new competitive advantage (but context sprawl is killing it)

So here’s what’s happening: The era of 6-month product launches is dead.

The underlying thing happening right now is that the notion of Tier 1, 2, 3, 4 launches is gone. Companies are shipping weekly now. Features that used to take months to bring to market are getting built and launched in weeks.

What I’m seeing: Companies that can’t adapt to this speed are getting left behind. Even HubSpot admitted “inbound alone won’t get you the same results.”

Here’s the actual bottleneck: If I tell you “Hey, we’re shipping this Tier 2 feature in 2 weeks, we’re launching broad market - you have 2 weeks to figure out positioning, messaging, and launch plan.”

Most PMMs would panic. They’re used to 3-6 month research cycles for launches.

The solution: You need systems for speed at every stage:

• Analysis super fast - Understand market positioning in days, not weeks

• Content creation super fast - Generate positioning and materials in hours

• Distribution super fast - Push updates across all channels immediately

The bottleneck: You need a growth engine running in the background, freeing you to focus on finding market gaps. But you can’t find gaps without deep product and buyer analysis.

3. The full-stack product marketer revolution

Here’s what I’m seeing happen to marketing roles in real-time:

Traditional marketing silos are getting absorbed into one powerhouse role - the full-stack Product Marketer.

What the future product marketer owns:

• End-to-End GTM Planning - From launch strategy to execution, working as the quarterback

• Competitive Intelligence - Not hiring specialists, building automated systems + deep analysis

• Customer Marketing - Handling success stories, case studies, buyer leaderboards directly

• Sales Enablement - Creating battle cards, objection handling, deal-specific content

• Content at Scale - Why hire a content marketer when you have the highest context about product + buyers + competitors? you can use existing tools to write programmatic SEO at scale. Use buyer calls to write blog content and LinkedIn posts for you. Like, all the blogs you see on this page is built on Cursor using public web data and reddit.

The core team structure I’m betting on with perfect balance of skills + taste:

Full-Stack Product Marketer - Highest context person, GTM quarterback

Growth Marketer - Creative experiments, viral stuff, performance optimization

Video Creator - Full-stack: recording, creation, editing, shipping

Brand Designer - Aesthetic differentiation, visual storytelling

Why this consolidation is inevitable: I can create “State of X” reports using buyer call data and public web scraping that’s higher quality than narrow specialists produce. I don’t need separate people for functions I can automate or handle myself.

It might sound harsh but: If you’re in a narrow in-house role (just competitive intelligence, just customer marketing), you’re going to face a hard time getting a job. The future belongs to PMMs with the highest amount of context - product knowledge + market intelligence + buyer psychology.

The difference between PMMs who thrive and those who get consolidated? Genuine curiosity about their craft. If you’re not obsessively diving deep into product architecture, buyer psychology, and market dynamics because you love it - you’re already behind.

4. Show real buyer wins (not fancy demos)

The perfect era for builder stories: Since everybody is becoming a builder, let those stories come out in the market.

Your buyer is the biggest way to win right now because most companies are creating fancy websites and demo interactive stuff that fails in production.

The shift: From demo-driven to product-led marketing.

What works:

• Show something that works in prod that somebody actually built

• Customer video stories: Real builders showing what they built

• In-production examples: Actual implementations, not demos

• Metric transparency: Real accuracy numbers, actual latency, honest limitations

  

Why this works: In the builder era, people want to see real examples of what’s possible. Your buyers become your best marketing when they can show off what they built.

Video and brand design will shine: You’ll work closely with video teams and brand design because brand recall matters more when everyone looks similar. You need your own style, your own narrative to differentiate from competitors. And the recent Sora 2 launched killed it.

Use the builder era: Customers want to show off what they built with your tool. This is your biggest marketing advantage:

• Feature buyers on leaderboards with real recognition

• Create industry-specific top performer charts

• Do showcase series featuring customer builds

• Flood social with customer videos showing actual implementations

5. Why marketers who can code will dominate

Here’s why I push for marketers who understand code:

You know the actual tech behind the tech product you’re selling. You need to feel that pain. If you can’t feel that pain, you can’t sell that product.

The AI era reality check: As I wrote in my previous newsletter about product positioning in the AI world - we’re in the instant gratification era. The TikTok-ification of B2B is real. People want to swipe, not schedule. They want to create images NOW. Build apps NOW. Generate code NOW.

The new reality:

• Old way: Homepage → Features list → “Book a demo” → Discovery call → POC → Legal review → Maybe buy in Q3

• New way: Homepage → Try it RIGHT NOW → Holy shit this works → Swipe credit card → Tell team on Slack

Why technical PMMs win this game: Marketers who can code will shine because they can:

• Build something customized for themselves using the product they are selling (feel the actual product pain, which means contextual positioning)

• Ship launches super fast using AI (match the pace of weekly product releases)

• Do things at scale by default (their natural builder instinct kicks in)

• Create the sandbox/playground experiences that buyers now expect

The AI-first rule: Scale + hyper-specific

The mindset shift: Change your thinking to AI-first, which means starting every marketing decision with: “How can I make this hyper-specific to individuals while doing it at massive scale?”

This is the unique power of AI that traditional marketing can’t match - super specific + massive scale without jeopardizing quality.

How to actually get there:

Start with one fun personal project using fun app building stuff like Lovable, Replit and if you are good enough then move to AI coding assistants like Cursor. You’ll fail for first 5 times, but nail it by the 6th time. Then pick your most annoying manual task and automate it - don’t ask permission, just build “ugly but working” solutions. v1 should embarrass you a bit.

Scale from there to: buyer call analysis → content generation → competitive monitoring → hyper-specific landing pages → custom outreach automation.

Your complete PMM tech stack

Essential tools for builder marketers:

Advanced automations in action:

• Customer Analysis → Landing Pages: here is my Clay setup for ABM which you can clone
  

• Competitive Intel: here is my github public repo and the video of how to use it. if you have replit/claude code/cursor - you can upload the code there and run it directly once you add the openai key locally in your system.
  

• Content Generation: here is my playbook that I created for writing these newsletters, reddit posts, linkedin posts, twitter posts, all using my thinking to industry insights flow.
  

6. Why technical depth is everything now

So here’s what’s going down: Kyle Poyar’s research shows monetization is shifting from charging for access to charging for work delivered. When buyers can test your product in 30 seconds and compare it to $25/month AI tools, surface-level marketing collapses.

Founders are noticing: “Very few product marketers understand the tech they’re selling”. There’s a massive shortage of PMMs who combine technical depth + marketing skills.

Generic vs. technical marketing:

❌ Generic: “AI-powered solution that streamlines sales workflows”

✅ Technical: Accuracy, benchmarks, latency numbers, model layers, product led proof from customer stories, public sandbox

❌ Generic: “Let’s start with benefits, not features”

✅ Technical: “Here’s deployment differences vs Netlify, database performance vs Firebase”

The PMM-engineering alignment framework:

1. Buyer-product fit:

• “This is buyer feedback - why is this feature important?”

2. Technical differentiation:

• “What’s our actual technical difference vs [competitor] that buyers care about?”

3. Positioning truth:

• “What’s the hardest technical problem we solved that competitors haven’t figured out?”

4. Proof points:

• “What technical metrics can we show publicly? (benchmarks, speed, accuracy, uptime)”

5. Buyer education:

• “How should we explain our technical approach to buyers who need to justify this purchase?”

6. Reality check:

• “Where are we technically behind competitors, and how do we position around that honestly?”

The context trinity: Your unfair advantage

Without all three contexts, you can’t do effective product marketing:

Product context: Use daily, understand technical architecture, know limitations, set up engineering/product calls asking “How should we explain our technical approach to buyers who need to justify this purchase?”

Competitor context: Track competitor moves (positioning, pricing, team hires, features), spot ecosystem gaps, find out shift in buyer behaviour

Buyer context: Analyze buyer calls for actual language used, understand decision processes (who influences/decides/implements), know technical/product objections for deal closures

Your action plan: Start now

My bet: Product Marketing + Growth skills = 🦄 . Without deep product context, nobody will hire you for senior marketing roles.

Don’t try to boil the ocean. Pick one manual task, automate it this week. Build on that success.

The goal: hyper-specific execution at massive scale. I have a couple of examples that I have already added about things that I have built. You may be able to use some examples from there, or you can start small by doing some automation for yourself.

Forward this to PMMs who are figuring out or are curious to figure out how to be ready for AI.

p.s. please answer the poll and thank you to all 100 subscribers. big milestone! 🙏💜

p.p.s. here is the waitlist form if you want to try www.qback.ai

Here's how to use it to not die in the AI era.

The companies actually winning? They're not talking about AI. They're letting you use it. Right now. On their homepage. Without talking to Brad from sales.

The proof:

• Lovable and Replit went beyond $100M in less than a year

• Clay = ABM today

• Mintlify ships features in 24hrs based on public tweet by a customer

They all did ONE thing: killed "talk to sales" and led with the product.
unless you are in deep tech security or healthcare.

And if you're still defending your enterprise sales motion from 2019, let me be super clear: you're cooked.

The death of "talk to sales" (finally)

Here's the brutal reality nobody wants to admit:

Every SaaS company built between 2016-2023 is built on a dying premise. The entire foundation was "book a demo" and "get a pricing quote."

That worked when buyers had patience. When evaluation cycles were 6 months. When committees made decisions.

But here's what changed:

The CEO is now the buyer. There's never been a push like this where the entire Board / VP / Manager, is demanding "get us AI or we're dead." Companies are literally pivoting, shutting down, or getting acquired purely because of AI.

Instant gratification won. We're in the TikTok era of B2B. People want to swipe, not schedule. They want to create images NOW. Build apps NOW. Generate code NOW.

Speed of building went insane. People are shipping features each week or day, month or quarter are only reserved for big stuff. Now, in all this hype, you got to treat everything as a launch.

Everyone who's actually building knows – the game completely changed.

Why Product-Led is the only AI Positioning

Let me break this down super simply because apparently all the LinkedIn thought leaders and so-called homepage consultants missed the memo:

Old way (you're screwed):

Homepage → Features list → "Book a demo" → Discovery call → POC → Pricing discussion → Legal review → Maybe buy in Q3

New way (you might survive):

Homepage → Try it RIGHT NOW → Holy sh*t this works → Swipe credit card → Tell team on slack or share on socials

The three categories of companies (and how screwed you are)

Category 1: 2016-2023 Enterprise SaaS

Screwed Level: 🔥🔥🔥🔥🔥

You have "Request Demo" as your main CTA.

But if you HAVE to save this, here's your only play:

• Build a sandbox. Today. Not next quarter.

• Put it on your homepage. Above the fold.

• Let people touch your product without talking to Brad from sales.

I don't care if your product is "complex." I don't care if you sell to Fortune 500. If Palantir and Anduril can show a product live or give trial account, you have no excuse.

Category 2: DevTools & Technical Products

Screwed Level: 🔥

You're probably fine. Devs never tolerated "talk to sales" anyway. Just make sure you have:

• Instant access and benchmarks

• Free tier that actually works or has generous credits to experience the product

• Docs that doesn't suck

Category 3: Post-2023 AI Startups

Screwed Level: You better lead with product or you're dead on arrival

If you're building AI and not letting people use it immediately, what are you even doing? Go look at Cursor. Look at Replit. Look at Claude. That’s your competition.

The 3 ways to show your product and ride any AI model hype

1. Replay the product

Record your product doing something sick. Show the replay on loop. Let people see the end result before they even sign up.

Or if you have workflow which gets triggered automatically then have a replay logic. Like exactly you see in ChatGPT or Manus.im

Example: manus.im/share/ir7GHEobMHq8rdW9M3TLp1?replay=1

2. Play with sandbox

Give people a playground where they can start building before they even create an account.

Example: rows.com

3. Let them remix

This is criminally underused. People don't know what prompts to use. They don't know where to start.

So give them templates. Remixes. Starting points. There has to be something in your category which are ideally templates or boilerplate stuff, find that and make it accessible for users.

You're in fintech? Create remixable financial models. You're in marketing? Remixable campaigns. You're in app building? Remixable apps.

4. Show legit benchmarks

Can't show the actual product? Show the scoreboard.

"We analyze calls 47x faster than [competitor]"
"Our AI writes code that passes 94% more tests"
"We process invoices in 12 seconds vs 3 days"

But have receipts. No vague "10x better" BS.

What This Means for Positioning

Forget everything you learned about positioning till 2024. Just stick to first principles.

New rules:

1. You ARE the role you replace or the workflow you displace:

"We're building an AI SDR" > "We help with sales automation" 
"We're your AI data engineer" > "We provide data solutions"
“We are a customer support platform” > “Building the next Zendesk killer”

2. Show, don't tell

Product demo/access on homepage > 1000 words of copy 
Live sandbox > 10 case studies
Public benchmarks or comparison > "Trust us we're good"

3. Ship your positioning

Your product updates ARE your positioning. Ship daily, position daily.

Where to start?

Step 1: Audit your current state

• Screenshot your homepage and circle every "talk to sales" CTA

• Count clicks to product access (if >1, you're losing buyers)

• Calculate your current demo-to-close time

• Find your biggest competitor's instant-access feature

Step 2: Pick ONE quick win

Based on your audit, choose the fastest impact:

• If no pricing shown: Add a calculator or "starts at"

• If no product access: Create one demo video replay

• If no templates: Build remixable or “can be copied” examples. even a spreadsheet/doc version of that would do.

Step 3: Get technical context

This week, ask your engineers:

• WHERE does our data come from? (understand the source)

• WHAT role are we actually replacing? (nail the positioning)

• HOW is our architecture different? (find the real differentiator)

• WHY does our approach matter technically? (explain the magic)

• “WHAT” is agentic in our product?

For founders:

If your first instinct is "but our product is different," you're part of the problem.
Every product can be experienced. Every. Single. One.

• Complex analytics platform? Show a dashboard with sample data.

• AI voice agent? Let people call a number RIGHT NOW.

• Enterprise workflow tool? Sandbox with pre-built workflows.

Hire marketers who understands how the tech genuinely works and for dev tools ONLY hire marketers who can build or code. all marketers will learn how to build / code but that is 4-5 years out. But, currently those who can will earn the top dollar (as it a scarcity)

The guerrilla product-led testing playbook

Can't get permission to kill "talk to sales"? Here are 4 tests you can run tomorrow to prove product-led positioning works (without breaking anything):

Test 1: A/B test “get started” pages

Set up two versions of your "Get Started" button:

• Version A: Lands on page that “looks” like your product console (same form, but styled like actual onboarding). You can even make the url be like console to make it real. Do mention in the form submission message that they will get the login details link in email within 1hr (else you are literally scamming)

• Version B: Standard demo request form

Or better, even you change the URL from
www.example.com/request-demo »»»» www.example.com/get-started (this would do the trick).

Track which one converts better. If people prefer the console, you've got proof they want to USE not TALK.

Test 2: Competitor page experiment

On your "vs Competitor" pages, test two button types:

• "Schedule Demo" vs "See It Working" (links to 60-second product video). No gated content ever.

Track which buttons actually get clicked.

Test 3: Email signature split test

• Have half your sales team use: "Book time with me [calendar]"

• Other half: "See our product in 60 seconds [Loom or whatever]"

See if it leads to anywhere. You can also add to every sales rep’s calendar.

Test 4: Free tool honeypot

Build a simple tool that does one small thing your product does. Put it on your website as "Free Tools."

The point: Every single one of these tests will show you that people prefer experiencing your product over hearing about it. And you can run all of them without asking permission or changing your main website.

Once you have data showing product-led approaches convert better, then you pitch the CEO.

Bottom line

We're in the instant gratification era. The TikTok-ification of SaaS.

Let your product do the talking. Not, your marketing deck.

And if you're still defending your "enterprise sales motion" from 2019, let me be super clear:

You can pivot to product-led positioning or you can watch your competitors eat your lunch while charging 1/10th your price.

There's no third option.

Restack and share this post if you love it, and share any feedback to make it better. And you want access to QBack tool, let me know.

You need to analyze competitors for a launch or do a feature gap analysis. Your options? Spend hours finding the right prompts for doing deep research on 3 different tools, figuring out pages to crawl and find the most recent data, pay $40k for some enterprise AI tool that gives you generic reports + where you'll have to do entire stuff manually or... build your own AI system that actually works.

I chose option 3.

I got frustrated with jumping through 10 different solutions on ChatGPT/Claude/Perplexity/Gemini, Deep Research in everything was very verbose, it was just not working out. Also they hallucinate as they miss out on context due to token limitations in the chat interface.

I generally do all my work inside Cursor, so I was like ok why not I do it here because I'm anyways using QBack to do buyer analysis, so I was looking for something which can help me do all the market analysis as well.

The real problem (that everyone's missing)

Here's what I discovered after testing every competitive intel tool out there:

Most tools focus on finding information. But the real bottleneck is understanding information in YOUR context. Honestly, that's where most tools fail.

That too it should be super dynamic. Like I want to scrape the list of customers of my competitors, tomorrow I want to analyze their docs page for strengths and weaknesses.

You don't need another dashboard showing you what's already publicly available. You need a system that understands YOUR product + competitors + buyer, tracks changes that actually matter to YOUR GTM strategy, and gives you insights you can actually use in YOUR next launch or battlecard or sales asset or whatever.

how I actually solved this (the hard way)

I hacked together a system that works. But it wasn't easy.

The First Attempt (that failed): I tried to do it entirely inside Cursor using Beautiful Soup plus a basic crawler. I picked one competitor to test with—Databricks. It had 876 pages under documentation and it just went bonkers. The system couldn't handle the scale and I wasted 8-9 hours maxing out my limit in Cursor.

The Second Attempt (also failed): I switched to Replit and built a basic solution there. That was too shitty. It just didn't work because what I'm trying to build is complex—a lot of steps, a lot of logic, a lot of saving stuff into memory. I wanted it to be fluid, like water. But it wasn't.

The Third Attempt (that worked): It took me 2-3 days of thinking about the architecture, then I was able to build it end-to-end in roughly 4-5 hours. Tested it in every shape and form, saved the data, ran multiple tests. Finally, something that actually works.

The biggest struggle? finding a scraping engine that could handle the huge load.

That was the biggest challenge.
and tbh, the Crawl4AI scraper did a kickass job. The max I tested was to scrape 140 pages in one go and it did not disappoint at all.

The Aha moment that made it work: When I created it as a very simple command line interface, it became super easy because the power was all under the user to input the prompts and everything.

The simple setup (that actually works)

Step 1: Input the URL You just give it a competitor's sitemap. The system sorts everything into product docs vs pricing vs homepage so you can pick what actually matters.

Step 2: Scraper magic I use Crawl4AI (an open source scraping engine that really helped me out). tbh, this was the hardest part to figure out.

Step 3: Everything stays in Cursor All the data gets saved right in your workspace. No external tools, no data leaks, no monthly subscriptions. Just pure, organized competitive intel. It's all yours.

Step 4: The analysis engine This is where it gets interesting. I just dump the scraped data and run it through GPT-5 mini which actually analyzes what each competitor does well, where they're weak, and what gaps exist in the market.

Step 5: copy-paste ready insights You get company overviews, product analysis, and feature mapping that you can literally copy-paste into slide decks or share with your team. It all depends on what you want to scrape and what info you're looking for.

I do everything inside Cursor because it's just a .MD file and I can do quick updates whenever I want.

See it in action + here is the GitHub link

Watch me scrape and analyze Anthropic's pricing page in real-time using this setup

In this demo, I show how the system scrapes 140+ URLs in minutes, analyzes pricing and features automatically, and generates actionable insights for under $0.10 - all without leaving Cursor.

Btw, if you want to use QBack or you want to do that similar level of buyer analysis that I've shown in this video or this kind of a setup, you can either:

1. Just fill out this form and you’ll automatically receive the access.

2. DM me over linkedin or reply to this newsletter. I'll give you the access so you can do your own stuff and do the buyer analysis faster.

What it does for me?

Feature Tracking Made Easy: I can now easily figure out which features are in beta, which are in GA, what's coming soon, and what's being deprecated. This is super useful for product positioning and competitive analysis.

Documentation Deep Dives: Since this is coming from documentation, I get detailed feature specs, limitations, and implementation details that you'd never find in marketing material or deep research as the game is all context. I can share this info with my product marketing team for launches and relay it to product and engineering teams.

Real-Time Updates: I can track changes over time and see what competitors are actually building vs. what they're saying they're building.

The difference?

My system gets context because I have all the raw data directly from the horse's mouth. So insanely lesser chance of any hallucination or outdated data.

What I'd Do Differently Next Time?

If I had to do this again, I'd give myself more time for thinking and figuring out the context. I didn't think about scale initially. I thought I could rate limit it and figure it out, but even doing rate limiting won't help because I'll have to update this and I'll max out on requests.

I also thought about using 6-7 APIs under freemium services and keep switching them, but that's also a no-brainer because I wanted to do everything inside Cursor. Plus, that's just annoying to manage.

I think I'd have to evaluate the technical aspects more thoroughly, especially for doing things at scale.

The real insight: If you're a product marketer evaluating AI tools in the market, look for tools which are dynamic and should give right bang for your buck. By default, compare everything with GPT, Gemini, whatever you have. It should give for one input, it should give 10 high quality outputs. And it should be very dynamic to your business needs.

I've already called this out in one of my LinkedIn posts.
That's where the future is. Nobody wants to either drop the prices of these siloed solutions (which only do one use case) so that it matches the current rate of other AI stack, or they need to give way more output at that price point (whatever they are quoting today).

This was more for fun for me, so that's why I did it. But I think if people or product marketers who are evaluating tools in the market and are thinking of something AI, they should look for tools which are dynamic and should give right bang for your buck.

What's next?

Try scraping one competitor's documentation this week. See how much insight you can extract with just a few basic scripts.

I'm curious: What's your current competitive intelligence setup? Hit reply and share your approach.

P.S. If this resonated with you, forward it to your product marketing/competitive intel/founder friends who are tired of overpriced tools that underdeliver. And you want access to QBack, let me know.

Every week, I'd sit down to write and think "there has to be a better way to do this."

So instead of just pushing through like most people would, I did what any PMM worth their salt would do: I built a system (at scale) to solve it.

Not just any system. A complete AI content creation workflow that takes my raw thoughts (which again I don’t type anymore) and turns them into newsletters, LinkedIn posts, and Twitter threads that sound exactly like me.

And I'm going to show you exactly how I built it - and give you everything you need to build it yourself.

What I Actually Built

Here's the setup I created:

Input: I just type my thinking into Cursor - doesn't matter if it's messy, stream-of-consciousness, or half-baked thoughts. it generally is my view on how I think it should be or how things

Process: My system analyzes my input, learns from my existing content, scrapes industry insights from all popular marketing and GTM leaders (Emily Kramer, Kyle Poyar, Kyle Coleman) and generates content in my authentic voice.

Output: Complete content package ready to publish across Reddit, LinkedIn, Twitter (X).

The entire workflow runs inside Cursor. No external tools (except browser automation to crawl LinkedIn), no switching between apps, no complicated setups.

I literally just input my thinking and it creates newsletters that write like me, think like me, and sound exactly like something I would actually publish (which I do though).

How the System Actually Works

Let me walk you through the real workflow:

Step 1: Raw Input Processing

I save my thoughts as a simple .txt file in my Cursor workspace. Could be 200 words, could be 1000 words. The system handles both.

Example: I literally just typed "I was being lazy around writing my newsletter" and the system expanded that into this entire piece you're reading.

Step 2: Industry Intelligence Gathering

The system automatically:

• Scrapes latest LinkedIn posts of (Emily Kramer, Kyle Poyar, Kyle Coleman) for marketing insights and viral patterns

• Analyzes top newsletters from Growth Unhinged, MKT1

• Extracts engagement triggers and high-performing content structures

• Updates my writing style based on my latest 32+ LinkedIn posts

Step 3: Content Generation

Using all that data, it generates:

• Newsletter: Expands my thinking to 2000-3000 words with my voice and structure

• LinkedIn Posts: 2-3 optimized posts with engagement predictions

• Twitter Thread: 5-8 tweets with proper threading and CTAs

Step 4: Cursor Integration

Everything stays in my development environment (all local). Generated content appears in organized folders which are pushed to Git. I can edit, refine, and publish directly from Cursor.

No switching between tools. No copy-pasting. Pure workflow efficiency.

The Technical Stack (For the PMMs Who Code)

This isn't just prompt or context engineering. It's a complete system:

Python Scripts: Automated content analysis and generation
Selenium Integration: LinkedIn scraping for (, , Kyle Coleman) insights
Feed Parsing: Newsletter analysis from top marketing sources
Style Learning: AI that adapts to my writing patterns over time
Cursor Workflow: Everything integrated into my daily coding environment

The system learns from:

• My existing LinkedIn posts (32+ posts analyzed for voice patterns)

• My newsletter content (automatically updated daily)

• My Twitter threads and engagement patterns

• High-performing content structures of others

• Industry leaders' viral hooks and frameworks

What Makes This Different from Generic AI Tools

Most AI content tools give you generic output that sounds like everyone else.

This system is trained specifically on my voice, my industry insights, and my PMM audience preferences. It doesn't just generate content - it generates MY content.

Example: When I input thinking about PMM automation, it knows to:

• Reference my QBack experience

• Structure content the way I naturally write

• Include real numbers and personal stories

• Talk directly to product marketers in my no-BS style

The result: Content that my audience can't tell was AI-generated because it sounds exactly like something I would write.

This is the perfect example of AI creating content about AI - but done authentically, with my voice, using real industry insights.

Send this URL to Cursor/Claude Code and ask it to copy it. That’s it! Literally use this setup to create something on your own.

Why This Matters for Every PMM

We're constantly creating content:

• Product positioning docs

• Competitive battle cards

• Launch messaging frameworks

• Demand gen campaign copy

• Thought leadership content for the CEO

This system handles the execution while you focus on stuff that actually helps you sell and win deals.

Most marketers are screwed because they only understand one piece of the puzzle.

Content creators: understand messaging but don't know product positioning.

Demand gen: understand funnels but don't know buyer psychology.

Growth marketers: understand metrics but don't know competitive dynamics.

I'm sharing everything:

This isn't a course (hate to see when people sell courses, don’t fall for that sh*t). It's the actual system I use daily, shared because I think every PMM should have access to this kind of leverage (and this should be free).

🏈 Quarter back your content creation the same way you quarter back deals.

Hit reply and let me know what content creation task you'd automate first.

The PMMs who figure this out early will have an insane advantage over those who wait.

P.S. This newsletter was generated using the exact system I described. I input my raw thinking about being lazy with newsletter writing, and the system expanded it into what you just read.

P.P.S. Follow along on LinkedIn @harshitjain7 or Twitter (@harshitqback) for updates.

btw, this offer is open, you can DM/comment/tag and i’ll try to build it out for you.

And Don (Head of Growth PMM at Vimeo), shared his usecase: Analyze NPS, produce NPS reports, and organize NPS comments by theme. 🧞‍♂️

His current pain: Just spend LOTS of time reading, analyzing, and organizing all those comments.

That’s how it began.

tbh, it took about 14 hour-ish end to end, but the real fight was a 4½hr flow-state sprint to ship it.

URL of my AI agent for NPS analysis (will deprecate the app after 15th July)

This is my 7th side-app, so a few patterns are finally sticking. I’m sharing the full play-by-play process and the “why I did it this way”—so you can copy the bits that work and skip the bits that don’t.

1. Nail the workflow before you even open an AI code editor / any vibe coding app

Every app I’ve shipped (this is #7) keeps proving the same rule: map the data journey first. I first use pen and paper to sketch the steps, and figure out, “What does the raw file look like? What does the final chart need? What moves in between?” Getting that chain clear and helps me refine my prompts ten times cleaner.

Btw, don’t fall for those stupid Linkedin posts about AI prompts. Just ask o3/opus4 to create prompts for you using all the best practices of prompt engineering. Most important piece is context about what and how you want the input and output to behave.

Once the flow felt solid, I forced myself to keep the stack thin just a browser UI since it is an internal app. You paste your own OpenAI key, run the report, delete the key when you’re done. I processed a 50 row sample CSV file multiple times for full executive reports while testing, and this was my spend ($0.99) for o3. So don’t worry much about token unless you are adding more than 1000 rows.

2. Code it piece by piece (and don’t sweat on the fancy UI yet)

UI can suck days if you let it. It’s super subjective and you can waste time just to refine edges, padding, boxes, colors, etc. Since this was an internal tool, a “good enough” is all I needed. The win was speed: ship something that cuts the team’s grunt work at 1/10 the old cost. Polish can wait till v2.

Here’s the order I followed:

A. Kickoff Prompt for Building:

Build a frontend-only AI analytics platform for customer survey data with these requirements:

ARCHITECTURE:
- React + TypeScript with Vite build system
- Frontend-first security (session-only API key storage, XOR encryption)
- Zero server-side data persistence for privacy
- Tiered analysis packages with transparent pricing

USER JOURNEY:
- Landing page with security transparency and trust indicators
- Drag-drop CSV upload with intelligent column auto-mapping
- Real-time AI processing with progress indicators
- Interactive dashboard with drag-drop widget customization
- Professional PDF export capturing all visualizations

AI INTEGRATION:
- Custom CX analyst prompts for theme extraction
- Sentiment analysis with business context
- Competitive intelligence from survey comments
- Revenue-focused strategic recommendations
- Dual AI provider support (OpenAI + Anthropic)

SECURITY FRAMEWORK:
- Prompt injection protection (40+ suspicious patterns)
- Rate limiting with browser fingerprinting
- Input sanitization and response validation
- Content Security Policy implementation

VISUALIZATION:
- NPS score distributions and trend analysis
- Sentiment breakdown with category clustering
- Theme modeling with interactive word clouds
- Competitive benchmarking with threat assessment
- Topic modeling heatmaps with hover insights

EXPORT CAPABILITIES:
- PDF reports with html2canvas chart capture
- CSV data export with company branding
- Shareable dashboard links
- Executive summary generation

B. Data-flow sketch:

Raw CSV Upload → Column Auto-Mapping → Data Validation & Preview → AI Analysis Pipeline → Dashboard Visualization → Export & Sharing

C. AI prompt layer for charts and visualizations:

You can add create n number of combinations to create custom dashboards but here are a few key ones that covers most of the use cases.

LEVEL 1: Basic Metrics
- NPS calculation (promoters - detractors)
- Response volume and completion rates
- Basic sentiment distribution

LEVEL 2: Theme Extraction
- Custom CX analyst prompts with few-shot learning
- Semantic theme normalization and clustering
- Frequency analysis with business context

LEVEL 3: Strategic Insights
- Revenue impact assessment and priority scoring
- Competitive positioning analysis from comments
- Implementation roadmap with timeline estimates

LEVEL 4: Executive Summary
- Key findings with business significance
- Strategic recommendations with impact metrics
- Action items with ownership and deadlines

D. Security Drill-down and User Experience Lessons:

• Preview data before processing - builds user confidence, you can double-check if the fields are mapped correctly.

• Real-time progress indicators - essential for long AI processes

• Fallback analysis - always provide some value even when AI fails

• Export capabilities - users need to share insights outside the tool. my export functionality didn’t do a good job. still figuring this out. but the dashboards are pretty lit, so you can take screenshots, record a loom video etc.

FRONTEND SECURITY LAYERS:
1. Input Validation
   - CSV file type and size validation
   - Malicious content detection
   - Path traversal prevention

2. API Key Protection
   - Session-only storage with auto-expiry
   - XOR encryption with random keys
   - Rate limiting per browser fingerprint

3. AI Prompt Security
   - Injection pattern detection (40+ patterns)
   - Content sanitization before AI processing
   - Response filtering for malicious content

4. Export Security
   - Filename sanitization
   - Content validation before PDF generation
   - Download path security checks

3. Why the flow matters?

I learned the hard way on app #4 that if you skip this step you’ll drown in prompt spaghetti later. Every prompt here feeds the next, classic chaining. Example: the “theme finder” prompt sends a JSON blob; that blob plugs straight into the “graph builder” prompt. No if-else rules, just the model deciding the next move from the last output. That tiny bit of reasoning is what turns a plain simple script into a baby AI agent.

4. Replit vs. Cursor: which felt better?

• Replit = if you are a marketer and have less engineering/zero then go for this. even if there is a knowledge gap in the how-to, you will still figure it out. basically you can easily build workable 1-2 step workflow apps from this.

• Cursor = this god mode. the prompting and way it thinks like an engineer is super powerful. might require a decent engineering knowledge or understanding to do some stuff, especially if you are going to save user info and have some complex backend logic. I’m still mastering my UI skills in this.

You can deploy using Vercel/Netlify. I used Netlify for this particular app.

Key Implementation Tips

1. Keep Security In Mind: Implement the security framework, the kick off prompt covers it

2. Build Incrementally: Start with basic NPS calculation, then add AI features

3. Test with Real Data: Avoid mock data from day one - use authentic CSV samples

4. Focus on UX: The user journey is crucial for right output as all the logic is sequential or parallel.

5. Plan for Scale: Structure components for easy addition of new chart types

6. Document Everything: Keep detailed notes on prompts you are using for each analysis and how the data flow is connected.

TL;DR
You can build your own AI NPS agent within a week. First sketch the data flow, keep the stack lean on Replit or Cursor, and ship on Netlify. Then tweak prompts till the findings make sense for your team. Have fun, keep shipping, and touch grass often ☘️

p.s. move on from frameworks and homepage copywriting BS you see on LinkedIn. dig deep into buyer and competitor data. it helps you build launch plans that break the status quo. find that gap/insight that helps you sell/market faster.

building has never been this fast. so PMMs need to step up and match that pace.

The job: talk only to people already shouting about the pain you fix.

People vent on socials every day or share what’s cooking in their company. Use those posts on LinkedIn, X, Reddit, Discord, or Slack like they’re live discovery calls.

Run your ABM with that context, and the rest (demo, signup, cash) follow. Miss that and every cold email is just more spam in the void.

Why “ABM” Is Mostly Noise

• Spray-and-pray tools abused this term way too much. People literally roll their eyes when you say it now.

• Job-change / funding / hiring alerts ≠ buying intent as they are super random. But, fresh pain = budget.

• Personalization circus like (AI avatars, fancy swag) eats budget but rarely lands a conversation.

Truth: Unless the target already feels the pain you solve, no cute token will move them. So design your ABM for conversation, not conversion.

[30-min video] how to set up your ABM flow with AI using Clay, Webflow, and Zapier.

1. Find the Pain Signal First

• Social listening > static lists. Track live vents on LinkedIn, Reddit, X, Discord, Slack, GitHub. You can use Zapier (reddit) or Clay to act as social listening tools for you.

• Keywords that matter: “any recs for…,” “stuck with…,” competitor rants. Use ChatGPT for channel-specific keywords (prompt below).

• Pull it all into Clay (or your tool of choice). Now you’ve got:

  • the company

  • the exact phrasing of the pain

  • the humans who shouted it

Batch ~200 accounts. Big enough to learn, small enough to stay personal.

Prompt for the keyword list goes here ↓

**System (role)**
You are a senior revenue-marketing analyst who specialises in intent-signal mining across LinkedIn, Reddit, X (Twitter), Discord, Slack communities and GitHub issues. Your goal is to return highly-targeted keyword patterns that reveal buyers experiencing pain your solution can solve.

**User (directive & context)**
Generate keyword *patterns* I should feed into Clay.com’s social-listening step to surface live conversations.  
Focus on phrases that express:  
• requests for recommendations (“any recs for…”, “looking for a tool that…”)  
• users blocked or frustrated (“stuck with…”, “can’t get X to work…”)  
• complaints / competitor rants (“<COMPETITOR> sucks”, “switching from <COMPETITOR>”)  

**Variables**  
• INDUSTRY = <INDUSTRY>  
• PRODUCT_CATEGORY = <PRODUCT CATEGORY>  
• TOOL_NAME = <YOUR PRODUCT / TOOL>  
• REGION = <TARGET REGION / GEO>  
• PERSONA = <BUYER PERSONA OR JOB TITLE(S)>

**Few-shot examples**
Q: INDUSTRY = Cloud Security; PRODUCT_CATEGORY = CSPM; TOOL_NAME = SkyGuard; REGION = North America; PERSONA = DevSecOps Lead  
A:  
• “any recs for cloud security posture tool”  
• “stuck with <competitor> cspm limits”  
• “is anyone using SkyGuard in prod?”  
• “moving off <competitor> and need options”  
• “DevSecOps folks in NA – best cspm 2025?”  

Q: INDUSTRY = E-commerce; PRODUCT_CATEGORY = Headless CMS; TOOL_NAME = ContentJet; REGION = EMEA; PERSONA = Digital Product Manager  
A:  
• “switching from <competitor> cms”  
• “stuck with monolith cms workflow”  
• “any recs for headless cms europe”  
• “product managers using ContentJet”  
• “<competitor> downtime rant”

**Task**
Using the variable values below, output:
 **Exact keyword list** – 30-50 phrases ready to paste into Clay (one per line).  
 

**Current variables**  
INDUSTRY = <YOUR INDUSTRY>  
PRODUCT_CATEGORY = <YOUR CATEGORY>  
TOOL_NAME = <YOUR TOOL>  
REGION = <YOUR REGION>  
PERSONA = <YOUR PERSONA>

**Output format (markdown)**
```markdown
### 1. Exact Phrases  
- phrase 1  
- phrase 2  

How to use this prompt:

1. **Fill in the five variables** for each product/market you monitor.

2. Run the prompt in ChatGPT (use o3 or above). Don't use older models like 4o.

3. Copy section 1 directly into Clay.com as the keyword list; section 2 into advanced filters; section 3 into manual search boxes on LinkedIn, Reddit, etc.

4. Repeat for more regions or personas.

Subscribe now

2. Map the Humans, Not Just the Logo

Enrich in Clay → target titles first. Directors, managers, senior VPs, CEOs, CTOs all work. Personally, I found CXO roles works for startups from Seed to Series D not beyond this as the CXO are super busy. I always lean on VPs, directors, and senior managers as these are the people who are tasked to solve the pain and have budget clarity or access.

3. Write like you’re talking to a real person

AI can help a lot if you feed it the right context. Use the exact words from their social posts and in every touch: emails, DMs, landing pages. Speak their lingo and show how your tool fixes what hurts.

Channels to hit

(pick 3-4, stay omnipresent and build trust by just showing up for next 4 weeks):

• Email sequence (Clay → any email sender). I run a 4-touch email flow using the dedicated landing page.

• LinkedIn DM + matched ads

• Light retargeting banners

• Optional: swag / gifting drop

Skip cold calls for dev tools, tbh I hate cold calls. If you must dial, fine, but at least let your human or AI SDR pull context from those social posts first.

Check out 👇 page created by Tim Davidson, to see some great ABM 1:1 ads inspo.

4. Build 200+ tailored pages without losing your weekend

Watch this 30-min step-by-step video to create a full Webflow × Clay setup to create personalized landing pages for every company.

First setup might take 15-20mins but once you are done, then it literally feels like freaking magic. As I created/updated 171 pages in 5 mins, in 1-click.

You can watch this on YouTube here.

5. What ABM actually wins you

ABM ≠ conversion. In B2B, where building is getting easier every day with AI and almost everything is PLG, your job is to hit the right persona with the right words so you can talk and build trust. Marketer Goal #1 is trust, not closed deals on day one.

If a prospect books a demo, starts a PLG trial, or just says “let’s chat,” you’re good. Conversions ride on the product now more than ever (everyone’s got a PLG angle. yep, even Palantir). ABM just gets the right eyeballs in the room.

Here is a screenshot (check the post) of how good ABM should feel like, it should feel like you got heard. 👇

TL;DR

• Listen for pain buyers share on social.

• Match it to the right persona/account.

• Show them a page in their own words with Clay × Webflow.

• Pick a few channels where your ICP hangs out, be everywhere, but keep each message hyper-specific to their post or the general problems they might be facing.

• Let the product close. ABM = conversations, not conversions.

Go ahead and give it a try. Forward this to your product marketing friends or founders who are pushing for ABM. Restack and share this post if you love it, and share any feedback to make it better.

✅ What do you do?
✅ How fast / secure / low-cost / accurate is it?
✅ How do you do it better?
✅ What’s your edge?
✅ How can I try it?

That’s it. That’s the dev checklist.

Developers aren’t reading. They’re parsing.

They’re like hackers who are scanning for mental shortcuts.
They don’t read. They parse.
They read your homepage like a GitHub README.

They hate fluff. Every extra word, every “unlock potential,” makes them work harder. That’s cognitive load, and the more you add, the less they trust you.

So, who are you actually writing for?

You’re writing for someone who:

• Spun up their own Postgres instance last night

• Just deployed a Lambda function from a café

• Has zero time to read a “Why Now” section

• Wants to test your product faster than you can explain it

• Lives in the command line / Cursor

• Judges your tool by how fast they can break it

• Loves games, memes (duh), maths, and has ADHD (who doesn’t in this world)
  

They are builders.
They don’t want promises.
They want instant, undeniable, usable proof.

Real example: Built a dev homepage last week

It’s going live this week. Here’s exactly how to do it with AI (obviously!)

Step 1: Analyze all buyer calls + competitor data

First, I used QBack (the one I built) to analyze Reddit, G2, webpages, and sales calls to pick all the insights. For calls you can go to your AI recording tool (Gong, Fathom) and run this prompt:

10x PMM Prompt Tip: Pull insights from your call recording tool

Remember, these are devs. They love games, maths, memes, and get distracted real fast. So your brand design and website structure have a huge role to play.

Step 2: Fix the website structure

There’s no rulebook here tbh. Especially for the ones you see on LinkedIn. Some dev websites just have one section, some feel like a README, and others are just one line with a pip install.

Fundamentally, your homepage should answer:

• What does this tool do?

• Can I use it today?

• How fast / secure / low-cost / accurate is it?

Do this: Pick 20–25 dev websites you love

🪦 Anything older than 2022 = fossil
📊 Add screenshots to a Google Slide deck
👀 Look for patterns, not templates

Then use your judgment as a PMM to go a bit wild but don’t forget who you’re writing for.
The pattern I found (so far):

• First headline and subtext says exactly what you do (hit both CTOs + devs if you can)

• Stats up front, as they love numbers

• Show “what’s under the hood” (make it fun, like game-like or command line)

• Show the product + how they can play with it

• Show how to try it now (like pip install)

• Must-have: a dev section that’s super technical

👉 Pro tip: Pick websites that match your company stage.
If you’re early (say Series A), don’t copy unicorns. Match energy, not scale.

There are also cool dev conferences where people nerd out hard, you can grab inspo from those too.

I picked 4 core sections to go deep on:

• Header

• Use cases

• Dev section

• How it works

These are the ones where devs sniff out BS the fastest.

Here are a few sample prompts that you can use to get started today:

10x PMM Prompt Tip: For all sections for a dev homepages

Step 3: Now let’s talk brand design

Either you should have design sense—or someone on your team does.

I’ve got my main man Gami. Been working with him for 5+ years.
He just gets behavioral psych and nails it visually.

Me? I suck at colors (they trip me out). So I let my team lead there.

Don’t do this solo. Get a few good brains in the room.
If you go full lone wolf, you’re cooked.

👉 Pro tip: Pick a few themes for your design: like ours, was FAST. And, show that in subtle ways. You can pick that from inspo from your core benefit/your brand name. We pulled inspo from (Video games, F1, Terminal UIs, TechCrunch, Dev Conferences like Data Council)

• Devs code, so we got that in our design language (dashes, grid, sharp edges)

• They love games (got our architecture to look not boring with snake effect for data transfer)

• They love maths, so numbers are everywhere

• They love old school/windows stuff effect like (not slick but 90s, so we got code fonts + retro UI touches)

Stakeholder Alignment Tips (don’t skip)

• Have one final approver (usually the CEO)

• Work with brand design. Relay your rationale to approver

• Don’t take design advice from people who aren’t designers

• Keep the team lean (3–4 people max)

• Be bold. Be memorable. Forget “safe”

• Share the why behind every decision. Let the approver know that even if you don’t agree with something or they don’t, it’s totally fine. You can always tweak it and get it to a place that works.

• Anyone quoting 2018 “enterprise software” vibes = red flag 🚩

Quick Recap: Skip the $10K Homepage Consultants

If you're thinking of dropping $10K on an agency (IYKYK) to write your homepage copy, pause. That move makes sense only if you have zero product marketing muscle in-house. Otherwise, write your own copy. You should know your product better than anyone.

Don’t overcomplicate it. Use AI as your quarterback at every step:

1. Go through your call recordings and competitors data.

2. List all developer-first tools launched post-2022.

3. Study websites related to your persona. Screenshot layouts, copy, CTA flows.

4. Treat each section as a modular block.

5. Spin your message around those blocks and tune it to your ICP and for your company brand style

Devs don’t care what your product promises. They care how fast it delivers. Let your homepage act like a README. That’s true for all builders.

Btw, everybody in the world will slowly become a builder / engineer (AI duh). So this homepage style will blend right with sales and marketers too. Go bold!

If this hits home, foward it to you PMM friends, dev founders, or whoever’s building.
Got feedback? DM me on LinkedIn or just reply below.

Be a 10x PMM. Use AI as your quarterback. Let’s go!

Why I'm Here?

Currently, I lead the product marketing team at e6data and was previously a founder (QBack.AI; closed Dec 2024). I’ve spent 9+ years driving GTM as the #1 product marketing hire aka founding PMM, launching products, running competitive intel, and creating messaging that actually sells for startups across different categories—Conversational Intelligence (Observe.AI), CRM (LeadSquared), Sales Enablement & Research (QBack.AI), Data Infrastructure (e6data).

That means I’ve marketed to everyone from hardcore tech nerds to folks who’ve never worked beyond spreadsheets.

4x/month, all things AI to solve daily GTM challenges product marketers and founders face:

• How to craft messaging when your product seems identical to competitors

• Building sales assets that reps love and actually use

• Launching new features for tech nerds or spreadsheet audience

Share it with founders and your B2B marketer friends