Lorenz Graef
3.4K posts

Lorenz Graef
@lgraef
Founder STARTPLATZ and AI Hub NRW, Chairman Golf Post, Deep in AI, Agent Coding
Cologne Beigetreten Ocak 2009
970 Folgt14.4K Follower

I commented on LinkedIn for 10h/day for 7 days straight.
541 comments from a single account.
This changed how we approach LinkedIn forever.
The results:
- 4,591 to 290,147 search appearances in 7 days (6,222% increase)
- 315,373 total impressions (92% came from comments)
- 136% increase in profile views
- +7.5% week-over-week follower growth
And here's what nobody tells you about LinkedIn commenting:
Your first week WILL feel like shouting into the void.
But if you quit before you understand position data? You'll never see what actually works.
Here's what the numbers taught us:
- Position 1: 1,570 avg impressions per comment
- Position 2: 875 avg impressions
- Position 5+: 404 avg impressions
Most people struggle to pass 500 impressions on their posts.
A single well-placed comment beats that.
But here's where it gets interesting...
The thing that worked? Not what we expected.
We started with long, thoughtful comments (what everyone recommends).
Turns out the algorithm - and readers - HATED it.
So we switched to 11-25 word punchy comments in position 1. Results jumped immediately.
We also discovered:
- Being 10 seconds faster = 5x more distribution
- LinkedIn notifications are delayed 2-5 mins - manually check Priority 1 creators at their posting time
- Replying to replies creates engagement loops that keep your comment visible longer
- Rotating 5 comment patterns prevents your profile from looking like a bot
Most people would call it quits after a tough start and pivot to posting instead.
We said: give us 7 days of real data. Because the data doesn't lie - it just takes volume to speak.
So we recorded the entire commenting system:
- The 3-tier creator priority framework (who to comment on and when)
- The exact daily protocol (morning, throughout day, end of day)
- The 5 comment patterns that generate replies (Data Drop, Reframe, Story, Contrarian, Addition)
- The speed setup guide so you never miss a Position 1 opportunity
- The tracking spreadsheet to measure impressions, profile views, and DMs by creator
- What NOT to do (we analyzed all 541 comments so you don't repeat our early mistakes)
This is the same system generating 300k+ weekly impressions from comments alone.
And converting at rates most people don't believe are possible from a free channel.
Want the complete playbook?
1. Follow me
2. Reply "POSITION"
I'll send you the full breakdown.
English

Claude Code + Nano Banana 2 is f*cking cracked 🤯
I built a skill inside Claude Code that writes JSON image prompts for Nano Banana 2, and the outputs look like they came from a professional photo shoot.
One plain-text prompt. Claude rewrites it as structured JSON with lighting, camera, composition, style, and negative prompts.
Then fires it off to Nano Banana 2.
All inside Claude Code.
Perfect for DTC brands and agencies who need high-volume ad creative without booking a shoot.
If you're using Nano Banana 2 for product shots and lifestyle images but every generation feels like pulling a slot machine lever — random lighting, inconsistent style, plastic skin, misspelled labels ...
This skill fixes the entire output:
→ You describe what you want in plain English
→ Claude rewrites it as a structured JSON prompt (lighting, camera angle, lens, depth of field, color grading — all of it)
→ Fires it to Nano Banana 2 via API
→ Saves the prompt + image in organized folders
→ You iterate on the style until it's dialed, then every output matches
No more slot machine prompting.
No more inconsistent brand imagery.
No more burning credits on unusable generations.
What you get:
- Photo-realistic product shots and lifestyle images on demand
- Full control over style, lighting, composition, and camera settings
- Saved JSON prompts you can reuse across every campaign
- A skill that gets smarter the more feedback you give it
Built 100% in Claude Code with a custom skill + Python scripts.
I put together a full playbook showing the exact skill, the JSON schema, and the workflow to set this up yourself.
Want the full playbook?
> Like this post
> Comment "BANANA"
And I'll send it over (must be following so I can DM)
English

I just vibe-coded a Meta ads tool that generates 100+ new static ad concepts in minutes 🤯
Pick a product, choose how many ads you want, and the AI creates 100+ unique concepts across completely different formats, not just remixes of one ad.
All inside Claude Code.
Perfect for DTC brands and agencies who are tired of spinning the same 3 ad formats and need net-new concepts at scale.
If you're spending hours briefing designers on new ad concepts, waiting days for a first round, getting back 4 options that all look the same, giving notes, waiting again...
And then doing it all over again next week because you burned through the last batch.
This tool eliminates the entire loop:
→ Ships with 17 pre-loaded concept formats (us vs. them, iMessage-style, bold billboards, authority figures, feature callouts, UGC product-in-hand, and more)
→ Each format has real ad examples baked in so the AI knows exactly what to generate
→ Select your product, load your customer avatars, pick how many ads you want
→ AI auto-selects the best concept mix for your product and audience
→ Preview the full concept plan before anything generates — remove or regenerate anything
No waiting on designers.
No recycling the same 3 templates.
No generating 20 ads that all look identical.
What you get:
→ 17 concept formats out of the box (disable any you don't want)
→ Upload your OWN ad formats with up to 10 examples — the AI learns and scales them
→ Every ad comes with the format, angle, target persona, hook, body copy, and full prompt
→ Multi-brand support with saved brand kits and customer profiles
I recorded a full walkthrough showing the V2 concept generator in action, and I'm sharing ALL the Claude Code prompts I used to build it so you can do it yourself.
Want the prompts + walkthrough for free?
> Like this post
> Comment "STATICS"
And I'll send it over (must be following so I can DM)
English

My OpenClaw bot runs 6 AI agents 24/7 - and I'm giving away the exact skill file for free:
- Scouts local businesses without a website
- Audits their web presence and scores opportunities
- Builds a custom demo site + walkthrough video automatically
- Sends cold outreach with the live preview + payment link
- Handles objections and closes the sale via email
- Maintains existing clients on autopilot
Bring your own API keys. No subscription. Fully automated.
Reply "SKILL" and I'll send it to you (must be following)
English

i made a 3-day Claude Cowork for Beginners course, and it's yours for free
by the end, you'll have a personalized AI teammate on your computer that:
• knows your style
• connects to your tools
• and produces finished work you can send immediately
here's what you get:
day 1: install cowork, set global instructions, and run your first real task (15 min)
day 2: workflows that replaced hours of my week, including building landing pages from a description and running full competitive analyses in one prompt
day 3: skills, plugins, and connectors so cowork actually knows how you work and can access your tools
+ copy-paste prompts so you can follow along as you read
like + comment "COWORK" and i'll DM it to you

English
Lorenz Graef retweetet

We were Clay's largest user at one point in time, hitting their platform 17.3 million times per week. Last month we replaced them entirely with a $200/mo Claude Code subscription.
I can't write code. Neither can James, my VP of Growth who built the replacement.
Here's the full story.
Clay is a GREAT product and I TRULY think most people should use it. But we hit their ceiling.
50,000 row limit per table.
12.5 million row cap per workspace.
Tables that take days to actually delete.
Clicking "run all" thousands of times and waiting days for things to clear out.
So
When you're processing millions of leads, all the above become the bottleneck of your entire business.
James had never touched Claude Code before. Three weeks after learning it, he built our entire core system.
With Clay, processing 1 million leads took 27 hours.
And it would error out often enough that we would always have to plan on hitting the “run all rows” button again on 20+ clay tables. IYKYK
but
Our new system waterfall enriches 1 million leads in 5 seconds. 272,000 leads PER SECOND.
AND On top of the core engine, we vibe coded a Google Maps scraper that pulls leads zip code by zip code across all 32,000 US zip codes.
AND An AI lead finder that hits 95% contact match rates where Apollo gives you about 30%.
AND Ad library scrapers for Google and LinkedIn.
AND An AI campaign analysis system.
AND An auto-refill system so clients never run out of leads mid-campaign.
One we started building with Claude, we just couldn’t stop
Now we have the data ready for clients sending 5 million emails a month within 1 week of signing the contract.
I put together the full system blueprint -- every tool, the tech stack, a Clay vs custom comparison, and a 6-step playbook for building your own. Plus a video walkthrough where I show you the live system and how each tool actually works.
Retweet or Reply CODE below and I'll DM it to you.
English

LinkedIn is the best platform for building your B2B business.
I’ve created a prompt that will turn your profile into a premium profile that attracts leads.
Watch your inbounds and positioning perception noticeably increase
Like + comment "Send" and I'll DM you the file.
(Must be following me)

English

Ok, this blew up!
I’ve got hundreds of questions in my DMs after this post.
So I launched an OpenClaw Community for founders.
Where you can get my full 11-agent setup, weekly lives, AMAs, and a community of founders like you!
Comment “Lab” to get the link.
Florian Darroman@floriandarroman
English

ClawCon NRW — Feb 18 at STARTPLATZ Cologne.
OpenClaw on phones, in companies, as your personal AI butler — whatever your setup, bring it and show us.
Lightning talks + demos + networking @ STARTPLATZ.
Builders across NRW welcome.
→ luma.com/8hxmrfak
English

🚨 JUST IN: @claudeai got a huge update today and I'm flattered to be a part of it. As of now, Claude Opus 4.6 can build mobile apps and prepare shipping them on the Apple + Google app stores.
We just launched Shipper a new package that empowers Claude to:
→ Build complete mobile apps
→ Recreate existing apps
→ Assure iOS & Android compatibility
→ Autofill listings for both app stores (app icon, images, descriptions, keywords, privacy policy etc)
Claude Opus 4.6 can do all of the above in one prompt for ~$0.17/app... Publishable from the first prompt & built in 5 mins, not months.
You can try it on Shipper by asking Claude to "create an AI chatbot like Perplexity" or "eBay style marketplace app for local goods".
To celebrate this huge step forward, if you comment "SHIP" you will get free credits.
English
Lorenz Graef retweetet

Windows users, rejoice -- you can now try Cowork
Claude@claudeai
Cowork is now available on Windows. We’re bringing full feature parity with MacOS: file access, multi-step task execution, plugins, and MCP connectors.
English
Lorenz Graef retweetet
Lorenz Graef retweetet

@coreyhainesco I already used the previous version and got very good results. Thanks for your great work.
English
Lorenz Graef retweetet
Lorenz Graef retweetet

CLAWDBOT CHANGED HOW I DO LINKEDIN
but not how most people think
everyone's using it to write content and automate posts
i'm using it to book sales calls
here's the difference:
most people: "write me a linkedin post about b2b sales"
me: "analyze this prospect's last 10 posts and give me 3 personalized dm angles"
most people: "generate 50 connection request messages"
me: "research this vp's company news and find a reason to reach out that isn't generic"
clawdbot isn't a content machine.
it's a research assistant that makes personalized outreach scalable.
the problem with linkedin outreach was never sending messages.
it was making each one feel human at scale.
clawdbot fixes that.
been using it to:
→ analyze prospect activity before dming
→ find talking points from their posts and comments
→ personalize at scale without sounding like ai
→ cut research time from 5 min to 30 seconds per prospect
put together "The Clawdbot LinkedIn Playbook":
- exact prompts i use to research prospects in seconds
- dm frameworks that get 25%+ reply rates
- how to personalize 50 messages/day without burnout
- the research workflow that books 20-40 calls/month
- real examples from campaigns to saas founders and pe firms
follow + comment "clawdbot" and i'll send it over
(repost for instant reply)
English


