Building a personal brand on LinkedIn is one of the highest leverage activities for a technical founder. But manual posting consumes your scarcest resource: time. This tutorial shows you how to build an AI founder personal brand setup that uses Claude as your writing partner and Make as your orchestrator. You will generate content ideas, draft posts, schedule them to LinkedIn, and manage engagement. All while keeping your authentic voice intact.

This is not a guide to generic AI slop. This is a guide to scaling your specific expertise. If you can read JSON and understand HTTP requests, you can build a system that posts for you while sounding exactly like you.

What You Will Build

You will build a fully automated personal brand pipeline:

  • Content Ideation Engine: Claude generates post ideas based on your past content and domain expertise.
  • Draft Generator: A Make workflow calls the Claude API, formats the response, and logs it for review.
  • LinkedIn Publisher: An authenticated API call posts the content at scheduled times.
  • Engagement Assistant: AI drafts replies to comments and DMs, queued for your approval.

By the end of this tutorial, you will have a live system that maintains your presence while you focus on building product.

Prerequisites

  1. Claude API Key: Sign up at Anthropic and generate an API key.
  2. Make Account: Create a free account at Make.
  3. LinkedIn Developer Account: Register an app at LinkedIn Developer and request the w_member_social permission.
  4. Basic API Knowledge: Familiarity with JSON and REST endpoints will help, but I will explain each step.

Step 1: Getting Started with Your AI Founder Personal Brand Setup

Why automate? Consistency compounds. Engineers love to over index on product and under index on distribution. A reliable AI founder personal brand setup flips that dynamic. You generate the ideas and the taste. The machine handles the repetition.

Start by testing your Claude API key. Open your terminal and run this command:

curl https://api.anthropic.com/v1/messages \
  -H "x-api-key: YOUR_CLAUDE_API_KEY" \
  -H "anthropic-version: 2023-06-01" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "claude",
    "max_tokens": 100,
    "messages": [{"role": "user", "content": "Say hello in 10 words."}]
  }'

If you see a JSON response with content, you are ready. Next, set up environment variables in Make. Create a new scenario and add an HTTP module. Use the same payload structure but make it dynamic. Store your API key in Make's Environment Variables feature under Settings. Do not hardcode credentials.

This foundation takes 15 minutes. Once it works, you have a secure bridge to the most powerful language model available. Now let us build the content engine on top of it.

Step 2: Content Ideation with Claude

The hardest part of posting is knowing what to say. AI is terrible at generating ideas from zero context. But it is excellent at pattern matching. Feed Claude your previous posts and let it generate variations.

Claude content prompts for founders must include three things: identity, examples, and constraints. Here is a template you can copy:

You are a CTO at a seed stage data engineering platform. You have built three companies before this one. Your tone is direct, opinionated, and avoids buzzwords. Here are your five best performing posts:

1. "Why I refuse to hire data engineers who don't know SQL."
2. "The biggest lie in data engineering: you need Kafka."
3. "We rewrote our pipeline in Rust and cut costs by 60 percent."
4. "Stop building dashboards. Start building data products."
5. "My co-founder said we should use a data lake. Here is why he was wrong."

Generate 10 new post ideas in the same style. Each idea should be a single sentence that makes a strong claim. Prioritize topics that challenge conventional wisdom.

Run this prompt through the same HTTP module you built in Step 1. The response will give you a list of angles. You select the winners. This system turns one hour of thinking into a month of content topics. Your job is curation, not creation.

I recommend storing the generated ideas in a Google Sheet or Notion database. Use Make's Google Sheets module to append each idea with a timestamp and a status column. Build a no code second brain with Claude and Notion to keep everything searchable.

Step 3: Automated Content Generation with Make

Now we turn an idea into a full post. This is where the Make AI content workflow shines. You will chain four modules together to create a repeatable publishing pipeline.

The Make Scenario Structure

  1. Scheduler: Set a daily trigger for 8 AM.
  2. HTTP Module: POST to Claude API with your chosen idea and the system prompt below.
  3. JSON Parser: Extract the title, body, and hashtags from Claude's response.
  4. Data Store: Log the full draft for your review.

Here is the system prompt you will send to Claude for the full draft generation:

Write a LinkedIn post in the voice of a technical founder.

Constraints:
- Use short paragraphs. Two to three sentences max.
- Open with a controversial opinion.
- Include one specific technical detail or metric.
- End with a question to drive engagement.
- Add 5 relevant hashtags.

Return the output as valid JSON with keys: "title", "body", "hashtags".

Topic: {insert idea here}

In Make, the HTTP module response will contain JSON. Use a JSON Parse module to extract the data. For example, the path data.content.0.text holds the string. Then use a Text Aggregator to assemble the final post string: 1. title + "\n\n" + 2. body + "\n\n" + 3. hashtags.

Why build this yourself instead of using a ready made tool? Because you own the logic. You control the prompt, the tone, and the output format. Off the shelf tools force you into their templates. Connect Claude AI to your favorite apps and build exactly what you need.

Step 4: Scheduling and Posting to LinkedIn

Posting manually is the bottleneck. LinkedIn API automated posting solves this, but it requires careful authentication. LinkedIn uses OAuth 2.0 with long lived tokens. Do not use third party proxies. Own your API integration.

Getting Your LinkedIn Access Token

  1. Create an app in the LinkedIn Developer Console.
  2. Request the w_member_social and r_liteprofile scopes.
  3. Generate an authorization URL and paste it in your browser.
  4. Exchange the code for an access token.
  5. Set the token to expire in 365 days using the expires_in parameter.

Once you have a token, add an HTTP module to your Make scenario. Set the method to POST and the URL to https://api.linkedin.com/v2/ugcPosts. Add a header for Authorization: Bearer YOUR_TOKEN.

Here is the payload structure for the LinkedIn API:

{
  "author": "urn:li:person:{your_linkedin_member_id}",
  "lifecycleState": "PUBLISHED",
  "specificContent": {
    "com.linkedin.ugc.ShareContent": {
      "shareCommentary": {
        "text": "{your_post_body}"
      },
      "shareMediaCategory": "NONE"
    }
  },
  "visibility": {
    "com.linkedin.ubuntu.GenericNetworkVisibility": "PUBLIC"
  }
}

You can find your LinkedIn member ID by calling the /v2/me endpoint with your token. Store this ID in a Make Data Store alongside your token. To see the full integration pattern, read our guide on automating business with Claude MCP for connecting Gmail, Sheets, and Slack.

Schedule this scenario to run every Monday, Wednesday, and Friday at 9 AM. The machine posts while you sleep. You wake up to engagement.

Step 5: Using AI for Engagement with Guardrails

Posting is only half the battle. AI LinkedIn engagement automation can multiply your reach, but it carries real risk. A bot that replies insensitively will damage your reputation fast.

Automating the First Comment

LinkedIn's algorithm favors posts that get early engagement. Use Make to listen for your own post via the LinkedIn API or a webhook. Then use Claude to draft a comment that asks a question or adds a supporting insight.

Example prompt for the first comment:

The user just published this post: {post_text}. Draft a comment that asks a thought provoking question or shares a counter perspective. Keep it to 3 sentences. Do not self promote.

Queue this draft in a Slack channel or Notion database for your approval. Do not auto post replies. The goal is speed with a human review loop. Automate customer support with AI agents to learn more about building safe human in the loop systems.

Reply Suggestions for DMs and Comments

When your post takes off, you will get flooded with comments and DMs. Create a Make scenario that reads LinkedIn notifications, passes the message to Claude, and generates three reply options. Set strict limits: no more than 20 automated suggestions per day. Never automate replies to negative or controversial comments.

Always respect the rate limits. LinkedIn's API will throttle you. More importantly, your audience will notice if you reply too fast or too generically. Taste is the moat. AI is a lever, not a creator.

Step 6: Tracking Results and Staying Authentic

Measuring personal brand automation means tracking the metrics that matter for a founder. Vanity metrics like likes are nice. But you really care about qualified inbound leads, speaking invitations, and partnership opportunities.

Extend your Make scenario to log post performance into a Google Sheet. Track these columns:

  • Post URL
  • Publish Date
  • Impressions
  • Engagement Rate (comments + likes + reposts / impressions)
  • Number of DMs received
  • Number of connection requests

Every month, feed your top three performing posts back into Claude as few shot examples. This retrains the model on what works for your specific audience. Your voice evolves. The AI evolves with it.

Authenticity is not about writing every word yourself. It is about owning the perspective. AI can write a post about why microservices are overrated. But only you can tell the story of the production outage that nearly killed your startup. Lead with your scars. Use AI to polish the framing.

Common Pitfalls

  • Pitching too hard: LinkedIn penalizes overt sales. Focus on valuable insights. The product pitch happens in DMs.
  • Ignoring the review loop: Never set and forget. Review every draft until you trust the system. Even then, spot check weekly.
  • Over automation: Engagement automation is dangerous. Always keep a human in the loop for replies.
  • Token expiration: LinkedIn tokens expire. Build a notification in Make that alerts you 30 days before expiry.
  • Generic voice: If your posts sound like they came from a generic AI, your audience will tune out. Invest time in your system prompts. Founder Branding in 2026: How to Use AI to Stand Out goes deeper into this.

Next Steps

Your personal brand is an asset that compounds over time. The system you built today gives you the consistency to keep investing, without sacrificing your engineering hours. Start with one post per week. Tune your prompts. Watch what resonates. Double down on the topics that start conversations.

Once this pipeline is stable, extend it. Add a Make scenario that monitors industry news and generates timely hot takes. Connect it to your CRM to track leads that came from LinkedIn. The infrastructure is modular. Your imagination is the only limit.

Cover photo by Yusuf P on Pexels.