Building a content engine, not a content calendar
Category: CMS / AI Workflow / Webshop & Learning Platform
I run a small webshop and a couple of online courses. Every year I used to build a big content calendar in a spreadsheet: color-coded rows, a date for every product drop and course update, all lined up with the kayaking season. Six weeks in, nobody looked at it, and I was still writing everything by hand the night before.
A calendar tells you what to publish and when. It doesn't write anything. What I needed was a content engine: a system that drafts, checks and distributes content, so I only do the parts that actually need me.
From one idea to seven pieces
An idea goes in: a new jacket model, a customer question I keep answering, a lesson I want to add to a course. The engine turns it into a draft, and I edit that draft once. Then it checks the text against pages that already rank on my sites, publishes it, and generates the spin-offs: category page text, social captions, email snippets, quiz questions, FAQ entries and a version for my other site.
On a good month I get about seven pieces out of every source. With the calendar I never got past 1:1.5, because I ran out of time and energy long before I ran out of ideas.
In the webshop
Product descriptions and category pages are no longer the bottleneck. When I add ten new items for spring, the engine drafts every description in my tone and reuses phrasing from last season's bestsellers. I fix the parts that matter. A seasonal campaign that used to take days now takes hours.
In the courses
I write one solid module summary. The engine pulls out the key points and turns them into quiz questions with answer options, short explainers and FAQ items in the course voice. Students get more material, and I don't write any of it from zero.
One backend for everything
It all runs on the same CMS that serves several sites and customer accounts. One backend, several content streams. When I change the engine, the change applies everywhere at once. That is why it scales: I'm not maintaining five tools and five calendars.
What it doesn't do
It doesn't replace the person doing the writing. I read and own everything that goes out. The AI is fast and consistent, but it has no judgment, and without someone in the loop the quality drifts and the voice fades, especially in my nerdy niche activities' lingo. The engine takes the repetition. The thinking stays with me.
Now scale it
I'm a small example. Ten new products a season is doable by hand, just painful. What if you add 100 products next spring? Or a thousand, each with sizes and colors? That is where a calendar gives up and an engine starts paying for itself. Oh, did I mention we actually create these engines?
How you can do this
You don't need a custom CMS to start today, but it sure helps. Just give us a shout! Pick one content type, get it working by hand, then automate the steps that hold up.
Pick one bottleneck. Choose the content you write most often and dread most: product descriptions, course summaries or the newsletter. Ignore everything else for now.
Give ideas one inbox. A spreadsheet tab, a Notion database or a field in your CMS. Each idea gets a row with the raw facts: supplier specs, the customer's exact question, the lesson outline.
Write a one-page voice guide. Who you write for, words you use and avoid, and three to five of your best existing pieces as examples. Every prompt gets this page. It is what keeps your voice from fading.
Make one reusable draft prompt. Input: the facts row, the voice guide and one example. Output: a fixed structure, such as headline, two short paragraphs and a spec list. Keep the prompt in a file, and change it whenever you fix the same thing twice.
List your spin-offs. Write down what each source should become, each with its own short prompt and length limit. For example: a 40-word category blurb, three social captions, one email paragraph, five quiz questions and three FAQ entries. The length of that list is your target reuse ratio.
Review once, then derive. Edit the main draft properly. Generate the spin-offs from the approved version, not the raw draft, so your fixes carry over. Then skim the spin-offs before they go out.
Check against what already works. Pull the search queries that bring you traffic today from Google Search Console and give them to the draft prompt, or run an SEO plugin such as Yoast or Rank Math before publishing.
Automate the hand-offs last. When the manual flow works, connect it with Zapier, Make or n8n: a new row triggers the AI, and the drafts land in your CMS as unpublished. Shopify, WooCommerce and WordPress all accept this through their APIs. Keep the publish button manual until you trust the output.
Measure three numbers each month. Sources in, pieces out and hours spent. If the ratio goes up but the hours don't go down, your review step is too heavy, and the fix is usually a better prompt.
Going from ten products to a thousand. Work in batches through CSV export and import instead of one product at a time. For variants, write one parent description and generate only the lines that differ, such as color, size or fit. Review everything at first. Once the error rate is low, keep full review for parent products and bestsellers, and spot-check the variants. You can do it! You can also contact us for help!