n8n Daily LinkedIn Auto-Poster: OpenAI Brand Feedback Loop Workflow

Scheduled trigger → fetch content ideas → AI draft post → scored feedback loop → auto-publish to LinkedIn

Who

Personal brand builders, social media managers, small marketing teams looking to automate LinkedIn content creation

You get

Automatically generate on-brand LinkedIn posts daily at a set time, optimize through an AI scoring feedback loop, and publish directly

You need

n8n · linkedin · notion

Steps

1. Set up schedule trigger and fetch content ideas

   Configure the Schedule Trigger node in n8n with your preferred daily time (default: 21:00). Connect an Execute Workflow node to call the sub-workflow for content ideas (workflowId: N5PyYK0EClSQRWNK). Ensure the sub-workflow returns data containing a suggestions array.

2. Split ideas and format AI input

   Use a Split Out node to break the suggestions array into individual items, then process them one by one with Split In Batches. Each idea is formatted by a Set node (Format For AI Input) into JSON: {"sessionId": "{{ $itemIndex }}", "action": "sendMessage", "chatInput": "{{ $('Split Ideas').item.json.suggestions }}"}, then passed to the AI Agent.

3. Generate content ideas from brand brief (optional sub-workflow)

   The template includes a separate Get _Content _Ideas branch: call the Brand Brief sub-workflow (workflowId: UbKh62LfKoMDeyUo), feed it into an OpenAI node (gpt-4o-mini) to generate 10 topic suggestions aligned with the brand brief. Output JSON format: {"suggestions": [...]}. Enable this only if the Brand Brief sub-workflow is available.

   Execution prompt · Source text。Get _Content _Ideas node messages.values[0].content field

   Copy prompt

   =You are a professional content writer. Create 10 topic suggestions that align with the provided brand brief.
   
   JSON output format:
   {
   "suggestions": [
   "Create a post about [...]",
   ...
   ]
   }
   Brand brief:{{ $json.content }}

4. AI Agent drafts post with feedback loop

   The core AI Agent node (Generate LinkedIn Post with AI) uses an OpenAI Chat Model (default gpt-4.1-mini) with Simple Memory. The agent calls two Tool Workflow nodes: Get_Brand_Brief to retrieve the brand brief, and Get_Content_Feedback to score the draft. The system message defines a 5-step loop: fetch brand brief → create post → get score → refine if below 0.8 → output final post.

   Execution prompt · Source text。Generate LinkedIn Post with AI node options.systemMessage field

   Copy prompt

   =
   You are a helpful content creator for Nabin Bhandari's personal brand. Use the below steps to create content.
   
   Steps:
   
   Always start by getting the brand brief using the Get_Brand_Brief tool.
   
   Create a post on the requested topic that aligns with the brand brief.
   
   Get feedback and a score on the post you created in step 2 by using the Get_Content_Feedback tool.
   
   If the score is below 0.8, use the feedback from step 3 to refine the post, then repeat steps 3 and 4.
   
   The final output should be the approved post.

   Execution prompt · Source text。Get Content Feedback node messages.values[0].content (system role) field

   Copy prompt

   =You are tasked with evaluating a piece of content written for Nabin Bhandari. Your role is to assess how well the post aligns with Nabin's personal brand, using the below brief as your guide.
   
   Please analyze the post based on the following criteria:
   
   Does the post have unwanted artifacts like "Let's try again" or "Here's a new post"?
   
   Does the content reflect Nabin's core values, voice, and style as described in the brief?
   
   Does it engage Nabin's target audience as described in the brief?
   
   Does it emphasize Nabin's expertise and domain authority as described in the brief?
   
   Is the tone, messaging, and approach consistent with the brief's objectives and positioning?
   
   After evaluation, output a score between 0 and 1, representing the alignment between the post and the brand brief. The score should be output in the following JSON format:
   {
     "description": "[short description here]",
     "score": [insert score here]
   }
   
   #Brand Brief:{{ $json.content }}

5. Publish to LinkedIn

   The final post is published via the LinkedIn node (Publish to Linkedin). Node parameters: text = {{ $json.output }}, person = Nabin Bhandari. You must configure a linkedInOAuth2Api credential in n8n's credential manager before activating this node.

6. Customization and notes

   Per template sticky notes: customize the AI Agent prompt to adjust tone, length, or style; ensure LinkedIn credentials are properly configured; change the trigger time in the Schedule Trigger node. The author recommends storing a clear brand brief in Notion and collecting at least 50 approved posts for OpenAI fine-tuning.

You are setting up an n8n workflow to automatically generate and publish LinkedIn posts daily. Please confirm the following prerequisites first:
1) Access to an n8n instance; 2) OpenAI API key; 3) LinkedIn OAuth2 credentials; 4) Notion API credentials (or alternative brand brief storage); 5) Availability of sub-workflows Get Brand Brief / Get Content Ideas / Get Content Feedback, or your replacement plan. Once confirmed, import the template, replace credential IDs, change the hard-coded author name (Nabin Bhandari) to your brand, and validate the AI Agent feedback loop with test data before enabling live publishing.