How To Build Your LinkedIn Content Engine with OpenClaw + Codex

I run multiple brands' LinkedIn content, outbound and GTM systems, and I don't write a single post manually anymore.

I don't use a ChatGPT prompt that produces the same corporate slop everyone else is publishing. I'm not staring at a cursor at 11pm trying to remember what I was supposed to post tomorrow.

I do this with Airtable + OpenClaw. Inside Airtable I have several tables, including Brand Context, source material, output drafts and content watchlists. Together, the system takes one idea and produces 28 posts a month that sound like the founder, position the brand as the authority, and feed straight into outbound.

This is the full breakdown.

By the end of this you'll have a LinkedIn engine you can build this weekend.

What you'll have when you're done:

An AI that understands your voice, not generic LinkedIn "thought leader" voice
A top, middle and bottom funnel content system that converts profile visits into booked calls
A hook bank that performs on LinkedIn specifically
28 posts a month with zero manual writing
A profile and content stack that nurtures every outbound conversation you start
An Airtable base that stores your sources, drafts, context and finished content
OpenClaw with Codex doing the content production work in the background

Let's build it.

What is a content engine and why Airtable plus OpenClaw replaces a manual content team

Many teams use AI for LinkedIn like this:

Open ChatGPT
Type "write me a LinkedIn post about [topic]"
Get back the same corporate slop everyone else is posting. Emojis, "in today's fast-paced world", three-word sentences pretending to be deep.
Spend 20 minutes editing it so it doesn't sound like a robot wrote it.
Post it, get 4 likes from your mum and a guy trying to sell you ghostwriting.
Repeat tomorrow.

That's not a system. That's just doing the work twice.

The problem is not AI.

The problem is that you're giving AI zero context about who you are, who you sell to, what your offer is, what your voice sounds like, or what good actually looks like on LinkedIn.

You're hiring a brilliant intern who has amnesia, then doing it again the next day, then again the day after that.

A content engine fixes this.

Think of it like onboarding a new content writer to your business. You could throw them in cold and pray. Or you could hand them a playbook that explains who you are, who you help, what your authority is, what your voice sounds like, what hooks perform on LinkedIn, what your funnel looks like, and what the goal of every post actually is.

That is what Brand Context does.

Except it is not sitting in a random Google Doc.

It lives inside Airtable, next to the source material, output drafts, profile watchlist, blog sources and rewritten content.

When you point OpenClaw with Codex at this Airtable setup and give it a topic, it doesn't just read one prompt.

It checks the context.

It checks the source material.

It checks what has already been used.

It builds a complete understanding of the brand, the audience, the offer and the funnel.

Then it writes the post back into Airtable.

The difference is insane.

Bad: a single ChatGPT prompt = a writer with no brief, no idea who you are, no idea what good looks like
Good: Airtable plus Brand Context plus OpenClaw with Codex = a writer who's been inside your business for two years and knows exactly what converts

This is what replaced the manual content writing process for me. I still review every post, but I'm not writing from scratch anymore.

One more thing before we start.

This is LinkedIn-only, not multi-platform.

LinkedIn is the platform where B2B founders, operators, agencies, consultants and commercial teams actually convert. Twitter is fun. Instagram is better for B2C. LinkedIn is where decision-makers spend money, so that's where I play.

This system is built for one job:

Build authority on LinkedIn that turns into booked calls.

Where to actually build it

Airtable plus OpenClaw.

That's it.

Airtable is the source of truth.

OpenClaw with Codex is the operator.

Brand Context is the voice, positioning and business brain.

The AI doesn't need a complicated folder tree to sound clever.

It needs the right context, the right sources, and a production workflow that keeps everything organised.

What to actually run the system through:

OpenClaw with Codex, which is what I use. You connect it to Airtable, give it the instruction set, and let it read the context, source material and output tables. Best setup if you want the system to become part of a bigger AI operating layer.

ChatGPT with custom GPTs also works, but it is more manual.

Cursor or Codex works for the technical guys who want the AI to read local files directly and update things over time.

Most people reading this should use OpenClaw with Codex.

It's the strongest setup if you're serious about turning content into a repeatable business system, because the same logic can connect to Airtable, Make, CRM, outbound, reporting and your wider GTM machine.

The Airtable structure

Here's the actual structure:

Blog Sources to Watch
LinkedIn Signals to Watch
LinkedIn Output Content
Outputted Blogs
Brand Context

5 tables. That's the whole engine.

Within Airtable, those tables are the system. Brand Context tells OpenClaw how the founder sounds and what the brand stands for. The source tables tell it what to learn from. The output tables tell it where drafts go for review.

Go build the structure right now.

Seriously, pause, create the tables in Airtable. Takes a few minutes.

What each table does:

Blog Sources to Watch is where you store websites, newsletters, blogs and source articles worth monitoring.

LinkedIn Signals to Watch is where you store the the type of signals you want OpenClaw to learn from. Not random influencers. Highly engaging content that overlaps with your market, audience, writing style, hooks or ideas.

LinkedIn Output Content is where the final LinkedIn drafts go. OpenClaw writes here. You review here. Approved posts move toward publishing from here.

Outputted Blogs is the same workflow for long-form content.

Brand Context is the DNA. Who you are, how you sound, what you sell, what words you use, what words you'd never use, what proof you have, what angles matter, what your outbound is doing, what your audience actually cares about.

This is what stops the content from sounding like every other founder on LinkedIn talking about "unlocking efficiency."

Now let's fill each part.

Table 1: Brand Context, the command center

This is the most important table in the whole system.

Every time OpenClaw writes or rewrites a post, it should start here.

If this table is generic, your content will be generic.

Spend the most time on this one.

Brand Context is NOT a dump of random company notes.

That's the mistake everyone makes.

It's a briefing that tells the AI who you are, what you're trying to accomplish, and how to execute.

The table should include:

Identity

Founder: [FOUNDER NAME]
Company: [CLIENT / BRAND]
Niche: [BE SPECIFIC, e.g. helping B2B companies build AI automation systems around their existing sales, marketing, content, CRM, reporting and operations stack]
Authority: [WHY ANYONE SHOULD LISTEN, e.g. systems built, clients helped, pipeline generated, posts written, workflows shipped, proof across the market]

The goal of every post we write:

Position the founder as THE authority in AI automation and GTM systems
Nurture every cold prospect already in our outbound system
Generate inbound leads that book calls without us asking
Make the brand synonymous with practical authority in its niche, not generic AI hype

We are NOT optimizing for likes, follows, or virality.

We are optimizing for booked calls, qualified conversations and commercial opportunities.

A few things about this table:

The Identity section is not filler.

The AI uses this to calibrate every word.

"Helping businesses automate work" produces completely different content than "helping commercial teams build AI sales, marketing, CRM and reporting infrastructure around the tools they already use."

Be specific, painfully specific.

The "we are NOT optimizing for likes" line is a forcing function.

It stops the AI from drifting toward viral content patterns and keeps it focused on the actual goal, which is pipeline.

Spend the most time on this table.

It sets the standard for everything else.

The execution instructions

This is what OpenClaw follows when creating content.

When given a topic or source post:

Check Brand Context. Would this topic resonate with our actual buyer? If not, kill it.
Read the voice, positioning, offer and audience notes.
Decide which funnel layer this post should hit: top, middle, or bottom.
Pick 2-3 hook options for the topic.
Pick the format that matches the layer and topic.
Write the post in Stob.AI voice.
Save it into LinkedIn Output Content as a draft.
Do not publish automatically.
End with the appropriate CTA based on funnel layer. Top = none or follow. Middle = comment for resource. Bottom = book call or lead magnet.

CRITICAL RULES:

No corporate jargon, no "leverage synergies", no "today's fast-paced world", no "I'm excited to share..."
No emojis unless ironic
No hashtags in the body. Max 3 at the end if any
First line is everything. LinkedIn truncates at around 210 chars, so if the hook isn't in line 1, the post is dead
Every post must sound like the founder voice-noted it to a friend, NOT like a thought leader giving a TED talk
Any post containing kill list patterns must be rewritten before publishing

The voice DNA

This is what stops your content from sounding like every other founder on LinkedIn.

The kill list is non-negotiable. Every output gets checked against it.

Core Personality

We're an operator who's been in the trenches.
Direct, unfiltered, no fluff.
We know our stuff because we've built the systems, not because we read a thread about AI agents.
We have opinions and we share them.
We don't pretend every AI trend matters.
We talk like we'd talk to a smart founder over coffee, not like we're giving a keynote.

Tone Markers

Direct and unfiltered. We say "most people are wrong" when most people are wrong
Strategic, not tactical-bro. We zoom out. We see systems
Slightly chaotic but purposeful. We interrupt our own thoughts when something more important comes up
Anti-guru. We hate "just use AI" and "10x your business overnight" energy and we're allergic to it
Specific numbers. "$4,200/month system" instead of "a great solution", "12,000 emails" instead of "consistent outbound"
Honest about wins AND losses. The losses are usually more interesting
Practical AI infrastructure over AI hype

Vocabulary

Words we use:

pipeline, leads, booked calls, funnel, outbound, ICP, authority, conversion, CRM, GTM, deal flow, cold email, LinkedIn outbound, sales machine, automation, orchestration, infrastructure, operating layer, workflow, handoff, reporting, source of truth, Make, Airtable, OpenClaw, Codex

Words we NEVER use:

moreover, furthermore, in conclusion
delve, synergy, holistic, game-changer, unprecedented, paradigm
"I'm excited to announce", "thrilled to share", "humbled to be"
"in today's fast-paced world", "needless to say"

Phrases we like:

"here's what kills me"
"most founders get this wrong"
"the actual reason is..."
"study this"
"your team is not slow, your system is"
"the work between the tools is where the money leaks"

Phrases that should never appear:

"without further ado"
"circle back"
"let's unpack this"
"deep dive"

Energy

Write like the founder is voice-noting a friend who asked for advice at 11pm, not like a thought leader giving a TED talk.

Allowed:

Start mid-thought: "ok so honestly..."
Interrupt yourself: "wait, actually the bigger problem is..."
Get heated. Use ALL CAPS for emphasis sparingly
End abruptly without a "key takeaway"
Be willing to sound a little unhinged in a good way

Killed on sight:

Clean story to lesson to takeaway arcs. Too polished
Wrapping up with a "remember:" or "the lesson here is:" line
Explaining things neutrally when we have an opinion
Sounding "professional" in the LinkedIn-thought-leader sense

KILL LIST, non-negotiable

If any of these appear in output, the post must be rewritten before publishing. No exceptions.

Bold text anywhere in the post body. All text is plain
Em dashes. Use periods, commas, or "..." instead
"Not AI-generated [anything]" - cut entirely
"Real X. Real Y." constructions. Just say the actual thing
"No X. No Y. No Z." as prose sentences. Only acceptable as stacked bullet lines
"No X. No Y. Just Z." or "Zero X. Zero Y." - kill every time
Single "No X" openers when a direct statement is cleaner
"X isn't Y. It's Z." construction. Sounds like a corporate deck. Rewrite directly
"Stop X. Start Y." in any form. Fortune cookie pattern, cut every time
"That's it." as a standalone closer
"Ask yourself the real question"
"Every single time"
"Here's what I learned:" as a transition
"Here's the thing about..." or "here's what most people don't realize..."
"Genuinely" / "honestly" / "truly" as filler
Fake-specific AI numbers. Use real client numbers only
Dropping subjects from sentence starts. "I realized it was wrong" not "Realized it was wrong"
Bumper-sticker aphorisms. Fortune cookie closings disguised as insight
"Real X" constructions. "real results", "real problem", "real work", "real value", "real impact". Just say the specific thing
Repeating the same point multiple ways. Say it once, move on
Over-fragmenting related clauses. Use connectors: but, and, so, because, which, when, while

This context gets refined over time.

As you write more, you'll notice patterns.

When OpenClaw does something well, write it down in Brand Context.

When OpenClaw does something cringe, kill that pattern in Brand Context.

Most of the magic of the system is here.

The platform rules don't matter if the voice is off.

Formatting + delivery

Capitalization

Lowercase by default for body text
ALL CAPS only for emphasis on 1-2 words at a time
Title Case never. we are not writing a textbook

Line Breaks

One thought per line
White space between every distinct thought
Never write dense paragraphs. Nobody reads dense paragraphs on LinkedIn, they scroll past

Length

Sweet spot: 1,300 to 2,000 characters
LinkedIn truncates at around 210 chars with a "see more" button. HOOK MUST BE IN LINE 1
Articles up to 3,000 chars only for deep tactical breakdowns

Hashtags

Body: never
End: max 3, only if relevant. zero is fine

Emojis

Default: none
Allowed: ironic use only, max 1 per post
Banned: anything that looks like motivational speaker energy

Links

Never in the post body. LinkedIn algorithm penalizes external links in main post
Always in the first comment with a "link in comments" reference

Paragraph Style

Short. 1-2 sentences max per paragraph
Numbered lists are fine, bullets work
"..." for trailing thoughts, but don't overuse it

What the post should look like on screen

A reader scrolls.

The first line stops them.

The first line must have one of: a number, a contrarian claim, a story hook, or a specific outcome.

Then white space.

Then the next thought.

If your post looks like a wall of text, rewrite it.

Who you're actually writing for

This is the context that stops you from writing posts for "everyone."

Everyone is no one.

ICP, Who We're Writing To

Demographic

Job title: founders, CEOs, CMOs, Heads of Sales, RevOps leaders, agency owners, GTM leaders and commercial operators
Company size: service businesses, agencies, B2B companies, SaaS companies, sales-led businesses and international teams with existing tools but weak process infrastructure
Geography: UK, US, Europe, Middle East and Asia
Industry: B2B services, agencies, events, AV, SaaS, e-commerce, AI, consulting, real estate, education, hospitality, automotive and companies with sales, marketing or operations complexity

Psychographic, what they actually care about

Scaling without adding headcount
Turning scattered tools into one operating system
Finding leverage. they don't have time to manually chase leads, update CRMs, write posts, follow up, report or copy-paste between tools
Looking like a credible operator to their board, team, clients or partners
Making AI practical, not theoretical
ROI from automation, not "AI transformation" with no business outcome

What keeps them up at night

"we have the tools but the process is still manual"
"our team is busy but I can't see what is actually happening"
"we need more pipeline but our follow-up is weak"
"we're hiring people to do work software should already be doing"
"we've tried automation before and it became a mess"
"everyone is talking about AI, but I don't know what to actually build"
"I don't want another SaaS tool, I need the tools connected properly"

What makes them stop scrolling

Specific numbers: "12,000 outbound emails", "$4,200/month the core offer", "10,000 leads routed", "30 prospects surfaced"
Contrarian takes that match what they secretly believe but never said
Tactical breakdowns of systems, not motivational fluff
Screenshots of workflows, Airtable bases, dashboards, Make scenarios, agent logic, CRM routing
Behind-the-scenes of operations they wish they had
Posts that show exactly where time, money or pipeline is leaking

What makes them keep scrolling, death

"Just use AI!" energy
Generic frameworks
Motivational quotes
AI-generated 3-word-line posts
"Here are 10 ways to automate your business" with zero substance per item
Content that treats AI like magic instead of infrastructure

The language that converts

Resonant Language, use these

systems, infrastructure, operating layer
pipeline, deal flow, qualified inbound
automation, orchestration, workflow, source of truth
ROI, conversion, revenue, visibility
faster follow-up, cleaner CRM, better reporting
sales machine, GTM engine, automated sales infrastructure
practical AI, AI operations, AI automation
premium positioning, differentiation, competitive edge
"actually" as in "what actually works"
specific numbers always, $4,200/month, 12,000 emails, 28 posts/month, 10,000 leads, 3-month rollout

Turn-Off Language, kill on sight

"viral", "growth hacks", "engagement pods"
"let's connect", "happy to chat", "open to opportunities"
"thought leader" even when accurate
"social media presence". we say "LinkedIn funnel" or "authority system"
"personal brand". we say "authority"
"AI transformation". too vague unless tied to actual workflows
"AI agency". we say "AI automation and systems orchestration company"

How to address them

"you", direct, peer-to-peer
Never "guys" or "folks" or "fam"
Coaching energy from someone who's been there, not above them

Outcomes they actually pay for

More booked calls per month from outbound and content
Faster response times
Cleaner CRM data
Fewer manual handoffs
Better visibility across sales, marketing and operations
Shorter sales cycles because content and outbound nurture before the call
Higher close rates because prospects already understand the offer
Less dependency on hiring more people
Systems that make a small team operate like a much bigger one

These outcomes should be in the subtext of every post, even when we're not selling.

Hooks, the most important engine part

Hooks decide 70%+ of LinkedIn post performance.

You can have the best post in the world, but if line 1 doesn't stop them, nobody reads it.

These are hooks that perform on LinkedIn specifically.

Twitter hooks don't work here, Instagram hooks don't work here.

LinkedIn has its own pattern.

Rule: First line must be visible BEFORE the "see more" cutoff.

LinkedIn truncates at around 210 characters.

The Authority Result Hook

"I've helped [N] [type of clients] [achieve outcome]. Here's the [N] rules I follow:"

Examples:

"I've helped teams build outbound systems that touch 12,000+ prospects a month. Here's what actually matters:"
"I've built AI automation systems across sales, content, CRM and reporting. Here's where most companies waste money:"
"I've worked across global marketing, AI automation and GTM systems. Here's the pattern I keep seeing:"

Best for: middle of funnel, building authority

The Specific Number Hook

"[Specific outcome] in [specific timeframe]. Here's what actually moved the needle:"

Examples:

"2,000 personalised outbound messages across email and LinkedIn. Here's what actually mattered:"
"30 prospects surfaced for one client from a system that replaced manual website crawling. Here's the structure:"
"10,000 leads routed through one intake system sounds simple until the agencies, forms and CRM rules get involved."

Best for: bottom of funnel, case studies

The Contrarian Hook

"You don't need [conventional advice]. You need [actual answer]."

Examples:

"You don't need more AI tools. You need the tools connected properly."
"You don't need another SDR. You need your follow-up process fixed."
"You don't need to rip out your stack. You need an operating layer around it."
"You don't need more content. You need content that warms up outbound."

Best for: top of funnel, opinion posts

The Pain-Point Hook

"[Describe the exact pain in their words]. Here's what's actually causing it:"

Examples:

"Your team is busy, your CRM is full, and nobody knows what is actually moving forward. Here's why:"
"You've got Apollo, Instantly, LinkedIn, Airtable and a CRM, but the process still feels manual. This is the gap:"
"Your outbound is not failing because of the copy. The follow-up layer is broken."

Best for: top + middle of funnel, attracts ICP

The Behind-the-Scenes Hook

"Most people don't talk about [thing]. We do it every day. Here's how it actually works:"

Examples:

"Most people don't talk about what happens between lead scraping and a booked call. We build that middle layer every day."
"Nobody talks about what actually breaks inside AI automation projects. It's usually not the AI."
"Most companies think automation means a workflow. The hard part is the operating logic around it."

Best for: middle of funnel, demonstrates depth

The Story Hook, for top of funnel

"[N] [time units] ago I [situation]. Today I [outcome]. What changed:"

Examples:

"A few years ago I was selling marketing services manually. Today I build systems that turn sales, content and CRM into one machine. The shift wasn't tools, it was architecture."
"I used to think AI automation was about saving time. Now I think it's about making small teams commercially dangerous."

Best for: top of funnel, attracts followers

Hooks We NEVER Write

"In today's fast-paced world..." bury this on sight
"I'm excited to announce..."
"Just got off a call with [name drop]..."
Any hook that starts with "I" as the literal first word unless the result is strong enough
"5 ways to [generic verb]..." too vague, no specificity
"A lesson I learned about..." the lesson needs to BE the hook
"Stop [X]. Start [Y]." anything. fortune cookie, banned
"Here's the thing about..." or "Here's what most people don't realize..."

How to Use This

Pick 2-3 hook formulas per topic.

Test them.

Track which ones get the most engagement from your actual ICP, not just total likes, likes from the right people.

Feed the winners back into Brand Context and Airtable.

The system that turns posts into pipeline

Every post hits one of three layers.

Mix the right ratio and your content drives leads.

Mix the wrong ratio and you build an audience that never converts.

Top of Funnel, TOFU - Attract

Goal: Get new eyeballs on your profile. Build the audience.

Post types:

Stories: how you got started, wins, losses
Shallow ideas / observations the ICP already agrees with
Personal life with a business angle
Hot takes on AI, automation, GTM, sales, agencies and operations
Behind-the-scenes of building the company

Goal of every TOFU post:

Get them to your PROFILE.

Frequency: 2 of every 7 posts, about 30%

NOT this layer's job:

Convert.

Don't try to sell here, you'll just look thirsty.

Middle of Funnel, MOFU - Demonstrate

Goal: Show you actually know what you're doing.

Demonstrate depth without giving away the entire service.

Post types:

Tactical breakdowns: here's HOW we do this thing
Frameworks YOU built
"Here's how we think about X" positioning posts
Industry observations that show you see the space differently
Objection handling: the things prospects worry about, addressed publicly
Breakdown of systems, workflows, tools and operating logic

Goal of every MOFU post:

Make the reader think "ok this person actually gets it" → click profile → eventually become a lead.

Frequency: 3 of every 7 posts, about 45%

This is the workhorse layer.

Most of your posts should be here.

Bottom of Funnel, BOFU - Convert

Goal: Turn readers into action takers.

DM you, book a call, download a lead magnet.

Post types:

Case studies with specific numbers
Lead magnet drops
"We've got capacity for X implementation slots this month" posts
Direct asks for the call
Breakdown of offers like the core offer, AI Ops System, GTM Infrastructure, Reporting Command Centre

Frequency: 2 of every 7 posts, about 30%

CTA on every BOFU post:

DM me [keyword]
Comment "[keyword]" and I'll send the doc
Link in comments to book

The lead magnet drop is the highest-leverage post type in the entire system.

People who comment for a free doc just told you they have the problem.

Now they're in your DMs.

Now you have a conversation.

The Weekly Mix

Posting 5x/week:

Mon: MOFU, tactical breakdown
Tue: MOFU, framework or how-we-do-it
Wed: BOFU, lead magnet OR case study
Thu: TOFU, story or hot take
Fri: BOFU, capacity post or direct CTA OR TOFU if you ran a BOFU on Wed

Posting 7x/week:

Mon: MOFU
Tue: BOFU, lead magnet drop
Wed: MOFU
Thu: TOFU, story
Fri: MOFU
Sat: BOFU, case study
Sun: TOFU, hot take or behind-the-scenes

What you DON'T do

7 BOFU posts in a row, you'll burn the audience
7 TOFU posts in a row, you'll grow followers who never convert
Random mix with no plan, that's how content stops driving pipeline

Post formats

Personal Narrative

Format: "I [did/learned/failed at something]" → lesson → takeaway
Best for: TOFU, stories
Length: 1,300-1,800 chars
Hook style: story hook

Tactical Listicle

Format: "[N] [things] for [outcome]:" + numbered items
Best for: MOFU
Length: 1,500-2,200 chars
Hook style: authority result hook

Case Study

Format: Problem → what we did → numbers
Best for: BOFU
Length: 1,500-2,500 chars
Hook style: specific number hook

Contrarian Take

Format: Bold statement → reasoning → reframe
Best for: TOFU, MOFU
Length: 1,000-1,500 chars
Hook style: contrarian hook

Lead Magnet Drop

Format: Hook → what's in the doc → "comment [keyword] to get it"
Best for: BOFU only
Length: 800-1,200 chars
Hook style: pain-point hook OR specific number hook

Carousel, LinkedIn document post

Format: 8-12 slides, one idea per slide, bold headlines
Best for: MOFU mostly, BOFU sometimes
Slide 1: hook, 8 words max
Last slide: CTA + your handle

The algorithm naturally promotes carousels.

Use them weekly.

Hot Take / Opinion

Format: Single observation → why most people are wrong → what's actually true
Best for: TOFU, MOFU
Length: 800-1,200 chars
Hook style: contrarian hook

Behind-the-Scenes

Format: "Here's what nobody talks about: [thing]" → what actually happens → what you learned
Best for: MOFU
Length: 1,200-1,800 chars
Hook style: behind-the-scenes hook

System Breakdown

Format: "The system looks simple from the outside" → tool stack → workflow → logic → business result
Best for: MOFU and BOFU
Length: 1,500-2,500 chars
Hook style: behind-the-scenes hook or specific number hook

Offer Breakdown

Format: Pain → why normal solutions fail → what the system includes → who it's for → CTA
Best for: BOFU
Length: 1,200-2,000 chars
Hook style: pain-point hook

Where ideas come from

Ideas are the bottleneck.

Solve this and the engine never stops.

Sit down to write without a swipe file → produce nothing
Sit down to write WITH a swipe file → produce 5 posts in an hour

Build the Swipe File

In this setup, the swipe file lives in Airtable.

That matters because the source material and the production workflow live in the same place.

Sales Call Transcripts

The single best source of post ideas.

Every objection a prospect raises is a post.

Every "I didn't realize that" moment is a post.

Every concern is a post.

After every sales call, pull out:

The 1-2 things they were most worried about
The 1-2 things that surprised them
The 1-2 things they said in their words that were better than how I would've said it

Each of those is a post.

Demos and Looms

Every demo creates content.

If you show someone a the core offer workflow, that is a post.

If you explain why Airtable is the source of truth, that is a post.

If you explain why Make handles the automation and OpenClaw with Codex handles the intelligence layer, that is a post.

DMs and Comments

Anyone who DMs you a question, that question is a post.

Anyone who comments with a strong opinion, that's a post.

Client Delivery

Every system you build is a content asset.

Examples:

Lead routing system
Outbound machine
CRM sync
Reporting command centre
AI content workflow
Form intake automation
Follow-up workflow
Reply classification
Sales dashboard
Invoice workflow

Each system can become:

a story
a breakdown
a diagram
a carousel
a case study
an objection handling post

Competitor Profiles

Not to copy, but to see what your ICP is engaging with.

If similar accounts are getting traction on a topic, your ICP is interested.

Take their angle and make it better.

Pain Points

List every pain your ICP feels.

Not 5, list 30.

Each pain is multiple posts.

One for awareness, one tactical, one case study.

Wins, Yours and Clients'

Every win is a case study post.

With specific numbers, with context, with what actually happened, not the polished version.

Losses

Every loss is a better post than most wins.

Deals that went cold.

Projects that got over-scoped.

Clients who misunderstood the value.

Follow-up that failed.

Systems that exposed a sales problem.

This is where the best content lives.

Industry Stupidity

Every dumb take you see in your industry is a hot take post.

You don't have to name names.

"Things I've stopped doing" / "Things I've started doing"

Endless format.

Always works.

The Weekly Workflow

Sunday, 30 minutes:

Ideate. Pull from the Airtable swipe file.
Decide on 5-7 topics for the week.
Tag each one TOFU/MOFU/BOFU.

Monday, 2 hours:

OpenClaw with Codex drafts all 5-7 posts.
It reads Brand Context.
It checks the source material.
It writes drafts into LinkedIn Output Content.
Review, tweak, schedule for the week using a scheduling tool.

Daily, 15 min:

Check engagement.
Reply to every comment in the first 30-60 minutes after a post goes live.
The algo reads engagement velocity and amplifies posts that get fast traction.

Friday, 15 min:

Review.
Which posts performed?
Which hooks worked?
What did the right people engage with?
ICP, not randoms.
Feed that back into Brand Context, Hooks and the swipe file.

Why batching matters

Most founders fail because they sit down to "write today's post" and they have nothing, so they post something mediocre, then they skip a day, then two, then they fall off entirely.

Batch writing eliminates the "what do I post today" moment.

Decision was made Sunday, content was written Monday, Tuesday-Sunday you just press publish.

This single change is what gets founders from "posting sometimes" to posting 5x/week consistently.

Because content drives traffic to your profile

Content gets people to your profile, and your profile either books the call or loses them.

Most founders spend 95% of their time on content and 0% on profile.

That's backwards.

Your LinkedIn profile is functioning as a landing page, not a resume.

Every section has one job:

Convert visitors into leads.

Profile Picture

Clear headshot, NO selfies
Professional doesn't mean expensive. Phone + window light works
Remove the busy background if needed
Smile. Look like a human someone would book a call with

Banner, most wasted real estate on LinkedIn

Communicates the OUTCOME you deliver
Include client logos if you have them
A specific number/result if you have one
Your logo present
Your brand colors

Example outcome:

"AI automation systems for sales, marketing, CRM, reporting and operations"

Or:

"Scale sales and operations without increasing headcount"

Headline, the W-A-R framework

W = What you do, specifically
A = Authority, proof element
R = Result, the outcome you deliver

Bad:

"Founder at [Company]"

This is a TITLE, not a positioning statement.

Good:

"Building AI automation systems for sales, marketing and operations | Founder of [Company] | Make, Airtable, OpenAI, OpenClaw, Codex"

Better:

"Helping businesses build AI sales, marketing and operations infrastructure around the tools they already use | Founder, [Company]"

Bio, 2 sentences max

Sentence 1:

Who you help and how.

Sentence 2:

Credibility anchor, numbers, clients, authority signal.

NOT your life story.

NOT your origin.

Save that for posts.

Featured Section, 3 items

Item 1: Your call booking link, for buyers ready now
Item 2: Your highest-performing post, for warm visitors who want proof
Item 3: Your best lead magnet, for cold visitors who want value first

Every type of visitor has a clear next step.

NOBODY hits a dead end on your profile.

Experience Section

Lead with your current company.

Format the description as a mini case study:

what you do, who for, what results you've gotten.

Past experience:

Keep only what's relevant to your current offer.

If your last role was unrelated, cut it.

Skills

List 1-2 skills max, the most relevant ones to your offer.

Get endorsements on those specifically.

Nobody cares if you're "endorsed" for 37 random skills.

The Test

Send your profile link to someone who's never met you.

Ask:

"in 30 seconds, can you tell what I do and who I help?"

If yes → profile works

If no → rewrite it

Content alone is slow

Content plus outbound is a machine.

Content alone takes 6-12 months to drive consistent inbound.

Outbound alone feels spammy and stops working at scale.

Combined, they compound.

Here's how they fit together:

The Loop

Content goes out → builds authority publicly
Outbound goes out → qualified prospects across email and LinkedIn
Prospect gets a message → checks profile
Profile is set up to convert
Prospect sees content already on the page = social proof BEFORE the conversation starts
Prospect replies, connects, clicks, books or keeps watching
Conversation starts. They already trust you, they already see your content, they've been nurtured before saying a word

What this means for content

Every post we write should pass this test:

"If a cold prospect lands on this profile after an outbound message, would this post help them trust me?"

If yes → publish

If no → kill or rewrite

This is why we don't optimize for likes or virality.

We optimize for:

Would my IDEAL CUSTOMER find this useful?
Would seeing this post move a deal forward?

What content does for outbound

Higher accept rates on connection requests
Faster reply rates in DMs
Better email conversion when prospects check the sender
Shorter sales cycles because prospects show up to calls pre-nurtured
Higher close rates because trust is already built
Better differentiation because you look like an operator, not another vendor

What outbound does for content

Targeted distribution. Your content gets seen by the EXACT people you sell to, not random scrollers
Engagement from your ICP, not random likes from your cousin's college roommate
Faster feedback loop on what works
More conversations, which create more content
More objections, which create better posts

You can't have one without the other and expect a working GTM system.

At this point you have the context, the Airtable base, and the OpenClaw workflow.

Now let's actually run it.

How to use the engine

Method 1: OpenClaw with Codex, what I recommend

Create the Airtable base.

Add the five tables:

Blog Sources to Watch
LinkedIn Signals to Watch
LinkedIn Output Content
Outputted Blogs
Brand Context

Fill Brand Context first.

Add source profiles and blog sources.

Start a new OpenClaw task.

Give it a topic and a layer:

Topic: Why most companies fail at AI automation even when they already have the right tools

Layer: Middle of funnel

Format: tactical breakdown / listicle

Instruction:

Follow Brand Context. Pick the best hook, write in founder voice, hit the layer goal, check output against the kill list, and save the draft into LinkedIn Output Content.

OpenClaw with Codex reads the context, checks the source material, writes the post, saves the draft.

Review, tweak, schedule, publish.

Method 2: Paste context

If you don't want to use OpenClaw with Codex, paste the relevant Brand Context into the chat.

Add the topic.

Give it the same instructions as above.

Works fine, just more manual.

Method 3: Cursor / Codex

For the technical guys:

Point an agent at the local folder or repo.

It can read AND write files.

Over time it updates hooks, refines tone rules, adds new pain points as they show up in your sales calls.

This is the more autonomous version.

The engine evolves itself.

What good output actually looks like

You give the engine this topic:

"The difference between AI automation that saves time vs AI automation that creates pipeline"

Bad output, ChatGPT with no context:

"In today's fast-paced digital landscape, AI automation has emerged as a powerful tool for businesses looking to streamline operations and increase efficiency. Here are 5 key differences..."

That's a 4-likes-from-your-mum post.

Good output, with the system:

most companies confuse automation with progress.

they connect a few tools, build a few workflows, save a few hours, then wonder why the business still feels the same.

here's the difference:

time-saving automation removes manual work.

pipeline-building automation changes what the team can commercially do.

these are not the same goal.

time-saving automation handles:

spreadsheet updates

copy-paste workflows

internal notifications

basic reporting

pipeline-building automation handles:

lead sourcing

signal detection

personalised outbound

follow-up logic

reply tracking

CRM movement

sales visibility

if your automation saves 3 hours a week but doesn't create more conversations, it's useful.

but it won't change the business.

the test:

would this workflow help the company find, contact, qualify, follow up with, or convert better prospects?

if no, it's admin automation.

if yes, it's GTM infrastructure.

only one of those becomes budget.

Same topic, completely different output.

The system isn't magic.

It's just context.

But context is everything.

What to do this weekend

Create the Airtable base.
5 tables, takes a few minutes.
Fill in Brand Context first.
This defines everything else.
Fill in the ICP, language, voice, kill list, offer, proof and outbound context.
Add LinkedIn Signals to find trending content within your niche.
Point OpenClaw with Codex at the workflow.
Give it a topic, write your first post, see what comes out.
Fix what's off, update the relevant context, write another post.

Don't expect the first version to be perfect.

The engine gets sharper every week as you feed back what's working.

What I'd avoid

Skipping the context table.

If you skip the context, your content sounds like everyone else's.

Writing for "everyone."

Everyone is no one.

Write for one specific person.

Treating it as a template.

The WHOLE point is the customization.

The connections between Airtable, Brand Context, OpenClaw and outbound are where the intelligence lives.

Ignoring the profile and outbound link.

Content alone doesn't convert.

Content plus profile plus outbound is the whole system.

That's the machine.

Enjoy.

Get Help: www.stob.ai

Get Content Machine: https://contentmachine.stob.ai/