Case study 04 · Content System · AI · Multi-Platform

The Content Automation System Behind Optimal Operations

One topic in. Full multi-platform content suite out. Built on Airtable, Make.com, Claude API, and a custom dashboard.

Why I built this

Most content systems for small businesses are spread across half a dozen tools and still need the founder to sit down for hours every week. I wanted the opposite. One central hub, AI doing the heavy lifting, multi-platform output, and about 30 seconds of human input per piece of content.

I built it for my own business first. It runs all of the Optimal Operations content now. Same system is what I package as a product for other small business owners.

What I built

Airtable as the content hub

Every piece of content lives in one Airtable base. Topic, target keyword, status, generated copy across every platform, image URLs, scheduled date, post performance. One record, full lifecycle.

Make.com as the trigger and generator

Make watches Airtable for a status change. When a record is marked Ready for Claude, Make fires a single Claude API call with a structured prompt and writes the full content suite back to Airtable as one JSON response.

Claude API doing the writing

One prompt, one call, returns the lot: carousel slides, Pinterest title and description, Instagram caption and hashtags, LinkedIn post and newsletter, blog title and outline, newsletter subject and body, SEO page title, meta description, slug, and image alt text. All written in my voice, with all my banned words excluded.

AI image generation

From the dashboard, one button kicks off AI image generation. A unique background image is generated for each carousel slide based on the topic, and brand-styled text is overlaid programmatically. Permanent URLs save back to Airtable. Every post has a unique visual, no stock photos repeating across the feed.

Custom Kanban dashboard

A Netlify-hosted dashboard that pulls live from Airtable. Content moves through columns: Idea, Ready for Claude, Generated, Approved to Post, Posted. Each card opens a preview panel with the copy, images, and a schedule picker. Edit, approve, schedule, all from one screen. Clients never touch Airtable.

GoHighLevel for publishing

Approved content with a scheduled date flows from the dashboard to GoHighLevel. GHL posts to Pinterest, Instagram, LinkedIn, the blog, and the newsletter on the scheduled date. Status updates to Posted automatically.

Analytics loop back to Airtable

Performance data on each post (reach, saves, engagement) flows back to Airtable against the original record. The next round of topic decisions is informed by what is actually working.

How it runs, step by step

  1. 1

    Input (about 30 seconds)

    • Type a topic and target keyword into Airtable
    • Set the status to Ready for Claude
  2. 2

    Automated generation

    • Make.com triggers
    • One Claude API call returns the full content suite
    • Airtable is populated (6 carousel slides, Pinterest, Instagram, LinkedIn, blog outline, newsletter, SEO fields)
    • Status set to Generated
  3. 3

    Dashboard

    • Open the dashboard
    • Generate images
    • Preview and edit
    • Pick a date
    • Status set to Approved to Post
  4. 4

    Publishing

    • GoHighLevel publishes on the date to Pinterest, Instagram, LinkedIn, the blog and the newsletter
  5. 5

    Analytics

    • Reach, saves and engagement flow back to Airtable and feed the next round of topics

The full process map

Input at the top, automated content generation in the middle, the dashboard review layer, then publishing fanning out across five platforms with analytics feeding back to the start.

content_system Content Automation System One topic in. Full multi-platform content suite out. cluster_input Input (about 30 seconds of human time) cluster_gen Automated content generation cluster_dash Dashboard (Kanban review, approve, schedule) cluster_pub Multi-platform publishing I1 Type topic and target keyword into Airtable I2 Set status to: Ready for Claude I1->I2 G1 Make.com triggers on Airtable status change I2->G1 G2 Single Claude API call Returns JSON with full content suite G1->G2 G3 Airtable populated: 6 carousel slides | Pinterest title, description, board Instagram caption + hashtags | LinkedIn post + newsletter Blog title + outline | Newsletter subject + body SEO page title, meta description, slug, alt text G2->G3 G4 Status set to: Generated G3->G4 D1 Open dashboard Content shows in Generated column G4->D1 D2 Click Generate Images AI image generation runs D1->D2 D3 Visuals built: Unique AI background per slide, text overlaid programmatically, permanent URLs saved D2->D3 D4 Preview content + images, edit copy if needed, pick schedule date D3->D4 D5 Status set to: Approved to Post D4->D5 P0 GoHighLevel publishes on scheduled date D5->P0 P1 Pinterest P0->P1 P2 Instagram P0->P2 P3 LinkedIn P0->P3 P4 Blog (WordPress) P0->P4 P5 Newsletter P0->P5 A1 Performance data flows back to Airtable (reach, saves, engagement) Feeds the next round of topic decisions P1->A1  metrics P2->A1 P3->A1

What this system delivers

30s

Human input per content piece (topic, keyword, status)

5

Platforms posted to from a single topic input

20+

Discrete pieces of copy generated per content topic

1

Dashboard to run the whole pipeline (no need to ever open Airtable)

Without it

Sit down at the start of every week. Brainstorm topics. Write the blog. Re-cut for Instagram, LinkedIn, Pinterest. Design slides in Canva. Schedule each one separately in each platform. Several hours per content piece across the platforms. Easy to skip a week when the work gets heavy.

With it

Type one topic and one keyword. Set status. Walk away. Come back to the dashboard, click Generate Images, preview, edit if needed, hit Approve and Schedule. The system writes for every platform, generates the visuals, queues the posts, publishes on the date, and tracks performance. Weekly content output becomes a 10 minute task.

How it's productised

The same system is offered to other small business owners across three tiers, depending on how much help they want with implementation.

DIY tier

Airtable base template plus a short video series walking through how to connect Make.com, the Claude API, and the dashboard inside their own accounts.

Done with you tier

Template plus a 90 minute live build session where the system is set up inside their accounts on a shared screen.

Done for you tier

Full build inside their accounts: Airtable, Make.com scenarios, Claude API, dashboard, GoHighLevel connection, brand voice document, tested and handed over with an SOP.

If your content output is suffering because the manual work to produce it just keeps stacking up, this is solvable. Book an Ops Review and we'll work out which tier fits.

Book an Ops Review

Back to all case studies

Want this kind of work in your business?

Book an Ops Review and we'll talk through what's breaking and what would actually fix it.