AI Content Production System
The Setup
The client was a B2B SaaS company with a small marketing team — two or three people, depending on the week — drowning in content production. They needed blog posts, social content, email newsletters, and product updates every month. The actual creation process took three to four weeks of grinding. By the time content published, it felt stale. The team was burned out. They were spending so much time making content that they had zero time for strategy.
I was brought in as fractional marketing director with a mandate: make the content engine run faster without hiring more people. The constraint was real. No new headcount. Limited budget. But the content calendar couldn’t slip.
The team wasn’t struggling with strategy or creativity. They were struggling with the blank page. With organizing their thoughts. With writing first drafts that took eight hours and then got rewritten anyway. With moving pieces between Google Docs, Slack, email, and their CMS. With keeping track of what was done, what was in progress, and what was published. They were drowning in logistics.
The Build
I built a system with three layers that stacked into one workflow:
Ideation layer: ChatGPT prompts trained on the brand’s voice, competitor content, and SEO gaps. In a single session, the AI generated a month of content ideas. Real ones. Not “write about AI” but specific angles, headlines, target keywords, and audience hooks. I spent a day upfront building the prompts — feeding in published content, brand voice docs, and competitor analysis until the output actually sounded like strategic recommendations.
Production layer: ChatGPT drafts with human editing. Not AI-generated content. AI-assisted drafts. The AI would produce a 1,200-word blog post outline with actual sentences. A human writer would shape it into the final voice, add examples, fact-check, and polish. The AI removed the blank-page paralysis. The human did the actual work.
Distribution layer: Zapier automations connecting ClickUp to the blog CMS, email platform, and social scheduler. Everything lived in ClickUp. Every piece of content had a card: Ideated → Drafted → Edited → Approved → Scheduled → Published. When a piece moved to “Approved,” Zapier triggered the distribution workflow automatically. The right content went to the right place. No manual copying and pasting.
The key insight was treating content like manufacturing, not art. Batch the ideation. Batch the drafting. Batch the editing. I call this “Making the Donuts” — do the same things, in the same order, every cycle. You’d be amazed what happens when you remove the decision-making from the daily work.
The Mess
The first month was rough. The AI-generated drafts were fine. Not terrible. Not great. The voice was off. The team had to spend almost as much time editing as they would’ve spent writing from scratch. I had to rebuild the prompts from scratch. I fed in real published content, brand voice docs, competitor examples, even transcripts of calls with customers until the drafts actually sounded like the company talking to its people.
The other mess was harder to solve: the team didn’t trust it. These were marketers who prided themselves on writing. Being told “here’s a draft from AI, now edit it” felt like an insult to some of them. I had to reframe the entire thing. “This isn’t replacing your writing. This is replacing the blank page. You’re still the editor. You’re still the voice. You’re still the quality control. You’re just starting from something instead of nothing.”
Once the prompts got better and the reframing stuck, the resistance dissolved. But I had to move through it. Most people don’t.
The Result
By month three, the system was producing a full month of content in a three-day sprint.
Day one: ideation and drafting. The team met for two hours. We talked through the month’s themes, customer needs, and competitive landscape. ChatGPT generated forty to fifty potential pieces. We picked the twenty that resonated. The AI drafted them overnight.
Day two: editing and approvals. The writers came in, shaped the drafts, fact-checked, added voice. Editors reviewed. Approvals moved pieces into the distribution queue.
Day three: scheduling and distribution. Zapier pushed everything to the calendar. Blog posts scheduled. Emails queued. Social content fed into the scheduler. Done.
The same team that was drowning was now finishing content production by Wednesday and spending the rest of the week on strategy, analysis, and creative experiments. Output increased — they went from eight pieces a month to twenty-plus without adding headcount. More importantly, the quality didn’t drop. The human editing layer kept the voice consistent. The AI just removed the paralysis.
The team went from resisting the system to protecting it. When a new hire suggested going back to the old way — “let’s just write blog posts like normal people” — the whole team shut it down immediately. They’d seen what it freed up.
The Takeaway
AI doesn’t replace writers. It replaces the part of writing that isn’t actually writing. Staring at a blank page. Organizing thoughts. Generating first drafts nobody loves. The real work is still human. The thinking. The editing. The voice. The judgment about what actually matters to your audience.
Once you separate those two layers, something weird happens: your writers get faster, not slower. They get more confident, not less. They see AI as a tool for their craft, not a threat to it.
How It’s Built
- Ideation: ChatGPT with custom prompts trained on brand voice, competitors, and customer research
- Production: ChatGPT drafts → human editing in ClickUp
- Project management: ClickUp with custom workflows and statuses
- Automation: Zapier connecting ClickUp → blog CMS → email platform → social scheduler
- Tech cost: About one hundred dollars a month (ChatGPT Pro + Zapier)
What Changed Because of This
This project became the template for every content system I built after. The “Making the Donuts” approach — batch everything, automate the handoffs, keep humans in the quality loop — is now the foundation of how I teach content production to my students and clients.
I learned that the bottleneck was never creativity. It was logistics. It was the overhead of managing moving pieces. Once I removed that, the team could actually think.
Related: Playbook: Weekly Content Calendar Generator | Playbook: One-to-Ten Content Repurposer