Every time we publish a post here, the same manual work follows: rewrite it as a LinkedIn post, sketch an infographic, remember to actually post it. We automated that — same dogfooding approach as our lead-intake agent, pointed at our own content pipeline this time.
Writing the blog post is the easy part. Turning it into a LinkedIn post in our voice, then briefing an infographic to go with it, is the step that quietly gets skipped when we're busy — which is most weeks. Nothing about that work needs a human doing the first draft; it just needs someone reviewing it before it goes out.
An n8n workflow, triggered by a webhook, that takes a blog post's title, URL, and content, then runs two things in parallel through Google Gemini: one call drafts a LinkedIn post in our specific voice — honest, no-hype, bullet points over buzzwords — and the other writes an image-generation prompt for a matching infographic and hands it to a dedicated Gemini image model to actually render. Both results get combined and emailed to us as a draft, never posted automatically.
Five steps, unattended until the review step:
The text side — the LinkedIn post draft — worked cleanly from the first real test. The image step took longer: getting an AI image model to reliably produce an on-brand infographic, within the limits of a standard API plan, took a few rounds of real-world testing to get right. That's a normal part of building something like this properly, and it's exactly the kind of groundwork we do before handing an automation off to a client rather than shipping the first version that technically runs.
This is now our standard post-publish step for every future blog post — including the one you're reading, which got the same treatment. It's a small, unglamorous automation, which is exactly the kind of task we think AI automation is best at.
Written by Kasa Tech. We build AI automation and custom software for startups and SMEs. About us