How to Automate Your Content Pipeline as an African Media Brand

You are publishing four times a week. You have two people. Every article goes through the same cycle: write, edit, format, schedule, share, repeat. Half the delay is not the writing. It is the hand-offs — the manual steps between draft and published that nobody has mapped or questioned.

This is the content pipeline problem. African media brands feel it harder than most. Budgets are tighter. Teams are smaller. Tools built for Western newsrooms assume infrastructure that does not exist here. AI is already reshaping how African businesses operate, and the media industry is no exception. We built a working solution for The Nairobi Post. This article shows you exactly what we built and how you can replicate it.

---

Why Your Pipeline Breaks Before the Article Is Even Written

The bottleneck is rarely talent. Most editors and writers in Nairobi are fast and capable. The bottleneck is coordination: who moves a piece from draft to published, which tool holds the asset, who updates the social schedule, who checks the format before it goes live.

When every step requires a human to physically hand work to the next person, your throughput is capped at your slowest step. On a two-person team, that slowest step is usually the same person doing three jobs at once.

The most common failure patterns we see across Nairobi media teams:

• Draft in Google Docs, manually copied into the CMS. Formatting breaks. Headers disappear. Someone fixes it by hand. That costs 20 minutes per article.
• Social captions written fresh every time. No template. No reuse. Every caption starts from a blank page.
• No publishing schedule. Posts go out when someone remembers to post, not when the audience is most active.
• No performance loop. Nobody checks what worked last week. Next week’s content plan is built on instinct.

Each of these is a workflow problem. None of them require hiring someone new to fix.

---

The Four Layers of a Working Content Pipeline

A functional content pipeline has four layers. You do not need to automate all four at once. Start with the layer that wastes the most time.

Layer 1: Intake and Assignment

This is where story ideas live and where assignments are tracked. Most small teams manage this in WhatsApp. That works until it stops — usually the moment you scale past three pieces a week.

Tool: Notion. Build one board with six columns: Idea, Assigned, In Draft, In Review, Scheduled, Published.

Automation: When a card moves to “Scheduled,” Make.com fires a notification to your social media scheduler automatically. One trigger, zero manual steps.

Layer 2: Draft to CMS

This step kills the most time. A Google Doc needs to reach your CMS — WordPress, Ghost, or Astro — with formatting intact. Nobody wants to reformat headers in a CMS at 11pm.

Tool: Make.com. Build a workflow that watches a specific Google Drive folder. When a doc is labelled “Ready for CMS,” the workflow pulls it, converts it to clean markdown, and creates a draft in your CMS.

We set this up for The Nairobi Post in under three hours. The editor never touches the CMS manually until the piece is ready to hit publish.

Cost: Make.com’s free tier covers 1,000 operations per month. A team publishing 20 articles a month will not exceed it.

Layer 3: Social Distribution

Every published article should produce at least three social assets: an X thread hook, a LinkedIn post, and a WhatsApp caption for your community groups. Writing these from scratch for each piece doubles your workload.

Automation: Chain a Claude or GPT prompt to your CMS publish trigger. When an article goes live, the workflow sends the title and opening paragraph to the AI, gets back three formatted captions, and drops them into a Notion approval table. A human reviews before anything goes out. The drafting step that used to take 15 minutes per article now takes 30 seconds.

Stack: Make.com + Claude API + Notion database.

Layer 4: Performance Feedback

The most neglected layer. Content goes out. Nobody systematically reviews what worked. The next week’s plan is guesswork.

Tool: Google Looker Studio (free). Connect your Google Analytics 4 account and build one dashboard: top 10 articles by pageviews this week, traffic sources, average time on page by content category.

Automation: Schedule a weekly email report from Looker Studio to your editorial team every Monday at 8am. The data lands before the planning meeting starts.

---

The Exact Stack Running on Our Nairobi Media Clients

Step	Tool	Monthly cost
Idea and assignment tracking	Notion	Free
Draft to CMS	Make.com	Free
Social caption generation	Claude API	KES 650–1,300
Social scheduling	Buffer	Free (3 channels)
Performance dashboard	Google Looker Studio	Free

Total: under KES 1,500 per month for a fully automated pipeline.

The only variable cost is the Claude API, and it scales with your publishing volume. A team publishing 20 articles a month spends less than KES 1,000 on AI caption generation.

---

Where to Start If You Have Nothing Built Yet

Pick the step that wastes the most time this week. Fix that one first.

For most media teams we work with, that step is draft to CMS. It is the most repetitive, the most error-prone, and the fastest to automate.

1. Create a Make.com account (free).
2. Set up a “Ready for CMS” folder in Google Drive.
3. Build a two-step scenario: watch the folder, push document to your CMS as a draft.
4. Run it for two weeks. Track the hours saved.
5. Add the social caption layer next.

The full four-layer pipeline takes most teams four to six weeks to build, working two to three hours per week. You do not need a developer. You need a workflow diagram and patience for the first setup.

---

What Changes When the Pipeline Works

A functioning content pipeline does one thing: it separates creative work from coordination work. Writers write. Editors edit. The system handles every step in between.

We have watched Nairobi-based media teams cut their publishing cycle from two days per article to under four hours. Not because they hired someone. Because they stopped doing manually what software does better.

If your pipeline is broken and you want it fixed, get in touch with Waziri Collective Labs. We do not sell strategy decks. We build the system and hand it over running.