Watch the video with chapters & transcript →
AUTOMATION
In short: A ready-made Make blueprint connects OpenAI to LinkedIn: import it, add your API key and organization ID, and the scenario writes and publishes posts automatically.
In this walkthrough, Ani Björkström sets up LinkedIn post automation using Make and ChatGPT: importing a ready-to-use blueprint, connecting it to OpenAI and LinkedIn, customizing the prompts for a specific brand, and scheduling posts to publish on autopilot — all without touching a single line of code.
In Make, open Scenarios, create a new scenario, and choose "import blueprint" — the tutorial names the scenario "LinkedIn automation". Then click the ChatGPT icon, add a connection, and enter your OpenAI API key and organization ID from the OpenAI platform; you also need money on your OpenAI billing account for the calls to work. The model is preset to GPT-4.1 mini, which the tutorial recommends leaving unchanged because it is quite good and cheap as well.
Every word wrapped in curly braces is a variable that must be updated; the rest of the prompt should be left as-is, since changes beyond the variables are not guaranteed to work. The system role defines who is writing and the assistant role defines what each post is about. The table below shows the variables with the tutorial's example values.
| Variable | Example value from the tutorial |
|---|---|
| Role | AI automation consultant |
| Industry | Finance |
| Tone / style | Friendly (professional or fun also possible) |
| Topic | AI automation |
| Target audience | Managers, developers, CTOs (line can be deleted) |
| Purpose of the post | Share news (or promote a product, provide a tip) |
| Tags | AI, AI automation, AI agents |
One nuance from the setup: choose the role you want to be seen as, not just your job title. The tutorial's author is a tech consultant, but selects "AI automation consultant" so the posts position her as an automation expert.
In the LinkedIn module, the only thing to add is the connection — pick "LinkedIn", not "OpenID Connect", because it is much easier to set up, then log in with your LinkedIn password and save. For scheduling, you can post at regular intervals such as every 15 minutes, but unless you want to bombard your audience, a day-of-week schedule is more sensible: the tutorial picks Tuesdays at 5:00 PM in the Europe/Stockholm time zone.
Save the scenario and click "run once" before trusting the schedule. A successful run shows a "1" on each module, and you can confirm the result by logging in to LinkedIn, opening your profile, and scrolling down to see the freshly published post — exactly what the tutorial does as its final check.
No. The entire setup is done by importing a blueprint, filling in connections, and editing prompt variables — no code is written at any point.
The blueprint comes preset with GPT-4.1 mini and the tutorial recommends keeping it, since it produces good posts at low cost.
You can, but the tutorial only guarantees the template works when you edit the variables in curly braces and leave the rest of the prompt text as defined.
Read the full video transcript, with chapters, on the video page →
Want this working inside your finance team?