In a recent LinkedIn post, Ruben Hassid explores a sophisticated method for leveraging AI to automate and enhance LinkedIn content creation. Hassid outlines a step-by-step process designed to transform a user’s existing successful content into a repeatable system for generating future posts, emphasizing a data-driven approach over mere aesthetic imitation.
Hassid begins by cautioning against feeding an AI model subpar content. Instead, he advises focusing on the top creators within one’s niche. “If your posts already crush it – great, train on them. But if they don’t, feeding Claude 400 mediocre posts just teaches it to be mediocre,” Hassid explains. He suggests identifying individuals whose LinkedIn presence is aspirational and using their content as a benchmark for the AI.
Data-Driven Content Analysis with AI
The core of Hassid’s strategy involves using AI, specifically Claude, to analyze successful LinkedIn posts and derive actionable insights. He details a process that begins with extracting public posts from a chosen high-performing profile using tools like Apify. This extraction, Hassid notes, is a safe method as it accesses publicly available data without logging into one’s own account.
Once the raw data is collected, Hassid stresses the importance of instructing the AI to perform precise calculations rather than relying on subjective interpretations. “Compute the exact stats in code. Don’t guess,” he instructs. This involves guiding Claude to correctly interpret data points often hidden within the raw export, such as media types and post dates embedded in activity IDs.
“The real format is in which media columns are. The date is inside each post’s activity ID. Claude can decode it – you never would.”
Hassid emphasizes that this meticulous data extraction and analysis are crucial for generating an accurate understanding of what constitutes successful content. The AI’s output, according to Hassid, should be a detailed report identifying winning strategies and a Standard Operating Procedure (SOP) for creating future posts.
Building a Repeatable Content Engine
The ultimate goal, as outlined by Hassid, is to turn these analytical outputs into a functional AI “skill” or template within platforms like Claude. This skill acts as a personalized content generator.
The Weekly Content Loop
Hassid describes the weekly workflow: invoking the custom skill, answering a few prompted questions about the desired post topic, and receiving multiple content angles, captions, and even suggestions for accompanying visuals. “One invocation = one post. Every week,” Hassid states, highlighting the efficiency of the system.
“Don’t run one viral recipe on repeat. A viral post is a burger – delicious, dangerous. Clone the best occasionally. Don’t eat it every day.”
He also provides crucial advice on refining the process. Hassid warns against over-reliance on a single successful formula, likening it to consuming a single dish daily. Variety, he suggests, is key to maintaining audience engagement.
Closing the Feedback Loop
A critical element of Hassid’s methodology is the continuous improvement of the AI model. “Close the loop. Every good post you publish becomes new training data. Better posts → smarter Claude → better posts. That’s the whole game,” he asserts. This iterative process ensures that the AI becomes increasingly adept at generating content that resonates with the user’s audience.
For those seeking a quicker implementation, Hassid offers his own pre-built AI skills, including extraction, recipe generation, and caption writing tools, freely available through his website. This allows users to adopt his exact system with minimal setup time.
📝 About This Content
This article is based on insights shared by Ruben Hassid on LinkedIn.
📅 Originally posted on July 21, 2026 | View original post on LinkedIn →